April 14, 2026 • 50 mins
In this episode, Priscilla is joined by Craig Stoss, VP of Solutions at Kodif, to break down what AI actually looks like in a real support workflow and how teams can use it in practical, low-risk ways.
They dig into why AI is more than just chatbots, where it actually saves time (and where it doesn’t), and how tools like summarization, tone refinement, translations, and reporting can make day-to-day support work more efficient without sacrificing quality.
Craig also explains the biggest misconceptions about AI, including why it can’t make human judgments, what it’s genuinely bad at, and why measuring success with traditional CSAT can lead teams in the wrong direction.
Plus, they discuss:
- Common mistakes when implementing AI
- How to think about a “frustration index” instead of satisfaction scores
- The skills support professionals should start building now
This conversation is for support teams who want to use AI without losing trust with their customers, and for anyone trying to figure out where AI fits into their workflow without overcomplicating things!
Where to find Craig:
We want to hear from you! Share your support stories and questions with us at happytohelp@buzzsprout.com!
To learn more about Buzzsprout visit Buzzsprout.com.
Thanks for listening!
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Priscilla (00:00):
Welcome to Happy to Help, a podcast about customer
support from the people ofBuzzsprout.
I'm your host, PriscillaBrooke.
Today we're getting practicalabout AI and customer support.
We'll talk about commonmisconceptions about AI,
actionable tips forincorporating AI into your daily
workflow, and what you need toknow to get the most out of your
AI strategies.
(00:20):
Thanks for joining us.
Let's get into it.
All right, so today is one ofthose episodes where I'm really
excited.
I know that at the end of thisepisode, everyone listening is
going to have some really goodtangible things to take and put
into their workflow tomorrow.
And so those kind of episodesalways get me really excited.
(00:42):
I feel like we've all listenedto podcasts where they give you
these big overarching ideas andconcepts to think about, which
can be super helpful.
But sometimes you just wantsomeone to tell you what to do
and how to handle something.
And so I'm really excited thatwe're going to talk about those
practical uses of AI, especiallyfor the people listening to
(01:03):
this who don't use AI much orare kind of scared of it or have
a smaller team and don't, youknow, have the funds or the
budget to buy a big AI platform,like how they can use AI
strategies in their day-to-day.
So joining us today to talkabout AI is Craig Stoss.
Craig is the VP of Solutions atKodif, which is an AI-driven
(01:24):
customer experience and supportautomation platform for
e-commerce customers.
He has over 20 years ofexperience in customer-facing
roles and is passionate aboutthe impact that customer
experience can have on yourbusiness.
Today, he's going to join usand bring us his insights into
AI and give us those practicaltips about using AI and customer
(01:45):
support.
So thanks for being here,Craig.
Craig (01:47):
Thank you for having me.
I'm so I'm so excited for thisconversation.
I think it's so relevant rightnow to so many different types
of businesses.
Priscilla (01:53):
Oh, yeah, for sure.
And I feel like it's one ofthose conversations that gets a
lot of people, whether they'resuper excited about AI or
they're super hesitant.
Like it's just there's a lot offeelings and emotions that come
along with AI.
And so it's always a goodconversation to have to like
break it down and look at it ina way that can kind of take that
emotion out of it.
We always kick off our episodeswith the same question.
(02:16):
So, Craig, who has made yourday recently?
Craig (02:19):
So my son and I went to get our haircut.
And the barber we have has akind of electronic waiting board
where you check in, you put youput your your name in.
But the way it works is whenthere's multiple people in one
group, you're supposed to selectthem all at once and hit go.
So you're kind of groupedtogether.
And I just forgot.
And then, of course, before Icould get my son checked in,
(02:41):
another person had checked in.
So now there was this big gapbetween me and my seven-year-old
son.
And I thought, oh great.
I'm gonna, you know, get myhe's gonna get his care haircut,
and then I'm going to have towait for 20 minutes and again
get my hair cut.
And he's gonna be sitting inthe, you know, getting bored.
And it was already about a 30,40 minute wait to get uh seen,
anyways.
(03:01):
Wow, yeah, it was a very busySunday at that the barber shop.
And I was I was reallyregretting this because when my
son gets bored, it's chaos.
And oh, and just to end theboot, they had a TV in the
lounge that the uh the TV wasnot working that particular day
on Sunday.
So that even made it worse.
He couldn't even be distractedby the TV.
Jordan (03:18):
That's unbearable.
Yeah.
Craig (03:20):
Yeah, yeah.
So I chatted with the woman whowas doing my hair and she was
very nice.
And and without a without meeven asking, she accommodated
and said, listen, we'll justreassign you.
And I thought, you know what?
That's a recognition of the thesituation that that you're in,
right?
And and the other person didn'teven notice, right?
You know, because my son's hairtakes two minutes to cut
(03:40):
because it's he's small.
So that made my day in a sensethat it was a really small thing
that, you know, bending therules a bit, but uh, but it
really it really helped probablyeveryone in the situation so
that my son didn't have to runaround the the barbershop going
crazy for 20 minutes.
Priscilla (03:57):
Yeah, it had it benefits and impacts that people
didn't even realize, right?
Craig (04:01):
Exactly.
Priscilla (04:02):
It's funny, I was talking about a similar
situation just this morning.
We were talking about refunds,and I was thinking about how
sometimes working in a situationin customer support, and it's
clear that a refund is the rightthing to do, whether they've
asked for it or not.
In the situation I was talkingabout this morning, someone had
upgraded to a paid plan, butthen immediately decided that it
(04:23):
wasn't gonna work for them andthis downgrade.
But they didn't ask for arefund and the specialist
offered one anyway.
It was one of those thingswhere the impact of the offer,
because it hadn't beenrequested, was even more than it
would have been had the personasked for it.
So if you had said, hey, canyou switch this around for us,
they probably would have done itand it would have been fine.
(04:44):
But the impact of the personnoticing that you needed that
without you having to ask forit, that's such a bigger impact.
And so I think it's one ofthose, like, as a customer
support specialist, being ableto identify when the right thing
to do something, even if it'snot requested, can be a cool way
to handle a situation.
And it shows like you careabout the situation more than
(05:07):
just reacting to the request.
I think that's a great story.
And shout out to them formaking that easier for everyone
involved.
Before we kind of get into AI,can you tell us ju"st a little
bit about how you got intocustomer support, but then also
a little bit about the workyou're doing now, kind of at
that intersection of AI andcustomer support?
Craig (05:26):
You know, I grew up in the 80s and and you know,
computers and things were justnot really a thing.
The internet hadn't beeninvented yet.
And so what I finally did getto touch my first computer, you
know, it at the time it wasn'tseen as it is today, where it's
like, you know, the internet andgames and flashiness, right?
It was it was a tool.
It was, it was, you know, justa fancy calculator.
(05:47):
And so, you know, I spent thefirst year coding and being so
fascinated that I could talk toan inanimate object and have it
do things that I asked it to do.
And as I kind of developed anddeveloped and developed that
kind of skill, and I got intoum, you know, there was a
program in Canada and ruralcommunities to bring when the
internet didn't go mainstream inthe 90s, there was a program in
(06:09):
in rural communities Canadawhere they would hire students
to teach senior citizens aboutwhat the internet was and how to
get an email address and how toget government forums.
And and I was selected to runthis kind of program.
And that was right there that Ireally saw this intersection
between technology and the humanparts.
And and from there I really didback down.
(06:29):
I did take a computer sciencedegree.
Uh, it was mainly in like youknow, the full life cycle of
software development.
It wasn't just development.
I learned how to writerequirements documents, I
learned how to QA, uh, I learnedhow to present ideas and gather
information from a customer.
All my internships during that,the during my university years
were in support departments orat least in customer-facing
(06:52):
departments.
And I just really felt likethat was where I wanted to be.
I didn't want to developfull-time, I wanted to kind of
be that intersection betweentechnology and people.
And over the years, I did I'vedone lots of different roles.
I was consultant for Fortune500 companies for software
development best practices.
I got into support leadershiplater after that.
Most recently running, uh, Iwas at a BPO and then uh running
(07:16):
some of their customer programsuh for their teams.
And then now in the AI world.
And the thing that's alwaysfascinated me about software in
general is support departmentshave all have this kind of like
cost center mentality.
And so software and softwaresolutions for support have
always been this kind ofcommodity in some ways, right?
(07:37):
It was one of these things thatthere's tons of cool solutions,
there's tons of systems.
10 years ago, when I first gotinto leadership, I was pitched
on tagging solutions andcategorization solutions and
reporting solutions andchatbots, early versions of
chatbots.
Yeah.
That stuff was just stuff thatyou would have a hard time
affording.
Yeah.
And what I where I think AI isdifferent, and what's really
(07:57):
cool about the AI industry andwhere we're going is that these
things are now more affordable.
The ROI, given the scale andscope of most AI solutions, is
much more clearer than say alegacy knowledge management
system.
You know, it it's very easy tosee that.
And so, yeah, that's how I gotinto it.
(08:19):
And and now that's how itintersects into my current role,
where I'm still in thattechnology and I'm still, you
know, as a solutions person, I'mstill talking to customers
every day about how they can usetechnology.
And that's ultimately mypassion is about making these
experiences high quality forthem.
Priscilla (08:34):
Yeah, I love that.
Can you tell us a little bitabout Kodif specifically and
what Kodif does and what itoffers for its customers?
Craig (08:40):
Yeah, I've known of CODA for about five years.
I met the founders at aconference and we got talking
about the future of AI, and itwas just on that cusp of Chat
GPT and open AI solutions comingout.
Priscilla (08:53):
Yeah.
Craig (08:53):
And one of the things that I've always believed in,
and this has nothing to do withAI at all, is that support needs
to be personalized andcontextual, always, right?
And CODA, when I was talking tothe founders, was one of the
first AI solution companies Italked to, and I've talked to
dozens at this point, thatreally was how it was built from
the ground up was everythingthey did, all their prompts, all
(09:15):
their back-end work was aboutmaking support contextual and
personalized.
So if you say something like,My grandfather lives in Germany
and I want to ship them yourproduct, it's not just gonna
say, Oh, I'm really sorry wedon't shift to Germany.
It's gonna say something like,Oh, what a great idea that you
want to ship something to yourgrandfather.
I hope he has a wonderfulbirthday, but I'm really sorry
(09:35):
we don't shift to Germany.
Wow.
You know, that's a much morehuman-sounding message.
It takes in the context yourquestion, it personalizes it,
those types of things.
And so every feature we buildhas like really strong abilities
to categorize, to understand,to empathize.
Like we have a lot of pet foodcompanies, and sometimes when
you're a pet food subscriptioncompany, people have to cancel
(09:57):
their subscription because a pethas died.
Priscilla (09:59):
Right.
Craig (10:00):
And you want to opt that empathy and say, Oh, I'm really
like that's not the time whereyou go in and be like, oh, well,
we'll offer you 25% discount ifwe if you were staying with us.
It's like, yeah, no, theperson's just lost a family
member, right?
So but that's what really drewme to Kodif initially, and and
that's something that goes intoevery feature we we produce.
Priscilla (10:18):
Wow.
Obviously, you're about to getinto all of this as far as you
know, how to use AI in a waythat is personal, is human, and
includes that empathy and stuff.
So let's jump into kind of likethe early days of AI and
customer support.
I remember when AI first waskind of coming on that it was
(10:40):
really scary, right?
And it was like, okay, is thisjust gonna like take away all
customer support roles?
And I had just really startedto solidify my existence in
customer support and just lovingit as work.
And then, oh no, AI is comingin.
Is this just gonna take it allout?
We've been working so hard tobuild this up.
And now are we gonna just giveit over to robots?
(11:00):
So I'm curious, you know,looking at how it's evolved over
the last three to four years,when you started working with
customer support teams, Craig,to implement these AI
strategies, what were some ofthe early reactions you were
hearing from customers and howhas that kind of changed over
the last couple of years?
Craig (11:19):
First and foremost, everyone's associated AI with
chatbot.
Those two terms were synonymousin many people's heads.
And I think the first thing wealways had to dispel was that's
not true.
I think that the most commoninterface for a AI tends to be a
chatbot, but that's not,they're not the same.
And the GPTs and the claws andthe uh and the Geminis of the
(11:43):
world are not anything relatedto the historic chatbots that
we're used to.
And and so that was the veryfirst thing was kind of getting
into this mode of why is thisdifferent or why is this better?
Because initial chatbotsrequired you to train it on kind
of questions and then answers.
(12:03):
So here's a list of questions,here's a list of answers, right?
Here's how they pair together.
And for the most part, therewas no kind of context of
synonyms, there was no contextof of different ways of asking
questions.
It was it was really, you know,match this keyword and then
this thing might be thisquestion, and here's some
answers.
And and you you remember thisfrom like early Zendesk AI days,
(12:25):
where like it would send you aresponse, and it was like these
three articles might help you.
And they were sometimes justway off base.
They just happen to mention thekeyword you used, right?
And that's not a shot at ZenDesk, right?
And I mean, that's thetechnology that we had.
Yeah.
But when when you talk tocustomers and that's what
they're associating with AI, youhave this upward battle of
saying, no, that's not whatwe're talking about.
(12:47):
We're talking about, you know,large language models, we're
talking about contextuallanguage models that understand
synonyms, they understand uhlanguage, they understand
colloquialisms, they understandthese things.
And I use the word understand abit loosely.
I think there's there'sprobably someone screaming at
their phone right now, theydon't understand anything.
And that's true.
Understand is a loose termhere, but they are trained on
(13:08):
this and they they under theyunderstand the connections, I
guess is the better way ofphrasing it.
Priscilla (13:12):
They can navigate it, yeah.
Craig (13:13):
Yeah.
So that's the first of theearly reactions.
I think how this has evolved isnow that in user-focused
industries, people have seen alittle bit about how AI is
improved.
They see these littlepersonalizations, they see that
the answers are not guessesanymore.
They're they're fairlyaccurate.
And so there's there's a littlebit more of that acceptance.
(13:35):
And when it's done well, andthere, you know, being able to
say something like, I would liketo speak to a human is a
seamless process to get to aperson or get to a human for a
certain type of ticket, or do sowithout you even, you know,
going back to the hairdresser,doing so without even asking,
just recognizing, oh, thisperson's frustrated.
Let's get them to a human.
Right.
That technology enables that.
And I think that that peopleare seeing that this is now a
(13:57):
benefit uh overall.
So there's definitely beenacceptance very recently.
Priscilla (14:02):
Do you see any current misunderstandings about
AI?
Is there like this far into it?
Are there things that you seekind of trends where people
misunderstand things about AIand how it comes into customer
support?
Craig (14:15):
Yeah, there are two big things that I would I would
argue uh I encounter on aregular basis.
The first is AI cannot makejudgment.
I think that's a reallyimportant distinction.
So, what we just talked aboutwith rules, AI has to have the
rules in place.
It can make judgments to apoint where you can teach it how
(14:39):
to make a judgment, but at thatpoint, it's not so much a
judgment as a different type ofrule, right?
And and so you know, I hearcomplaints all the time.
I was actually, we're going ona trip with some other couples
in a in a couple weeks, and wewere talking about because we
all have children, and they weresaying, well, every airline
seemed to have a differentalgorithm to assign seats with
adults next to children, andwhich is true.
(15:00):
It's interesting because Iwould argue that a lot of that's
done using some version of AI.
There's some algorithm inthere.
It may not be traditional LLMAI as we're talking about, but
there's some algorithm in there.
And it's like, well, they'reclearly just not putting,
they're either purposely notputting the rules in correctly
to make you buy a seat and forceyou to, or the rules are wrong,
(15:21):
right?
Because that's something Ithink you can define.
So going back to the judgmentpiece, so if I say to it, here's
how to judge if someonedeserves a refund, and then say
things like, you know, if theorder is greater than 30 days
old, they cannot get a refund.
Well, that that's a hard rule.
Cannot get a refund.
Right.
Let's use the dead pet example,right?
Like if the customer expressedtheir pet passed away, well,
(15:44):
maybe we'll make an exception.
You can you're allowed to makean exception for that or
something, but you have to teachit how to make that judgment.
You can't make a human-basedempathetic judgment.
And that's a big, big one thatcustomers ask for all the time,
especially when it comes tothings like warranties, where
it's like, well, you know, youhave to make a judgment.
Was this a manufacturer'sdefect or was this a uh was this
human error?
(16:04):
I mean, I don't think humans doa great job at that, let alone
trying to teach an AI how to dosomething like that.
So that's the first one.
And then the second one is thatthere are certain things, and
this will change, and itprobably will change by the time
you release this episode, butthere are certain things that AI
is just not good at.
I saw a LinkedIn post wheresomeone asked an AI to draw a
(16:26):
map of Europe.
Uh and by AI I mean a largelanguage model, and it did a
horrible job.
It lay it was two differentFrances, uh, it labeled Berlin
instead of Germany.
Like it was just, it was awful.
It was a horrible map, right?
And he was like, checkmate AIpeople, you know, can't even
draw a map of Europe.
And I was like, but that's youknow, that phrase, I think it
was an uh Einstein phrase wherehe said, if you judge a fish by
(16:48):
its ability to climb a tree,it's gonna fail.
That's that's what this isdoing, right?
You're judging an AI based onsomething it wasn't intended to
do.
And so when we get customers orwhen I have conversations with
people about AI, that'ssomething that I always am very
specific of.
This is a large language model,and like the key is in the term
(17:08):
language, there, right?
It it's very good at guessingthe next word based on the
previous 10 words.
It's not good at certain otherthings and it will improve.
And you can see some of theproblems with the improved AI uh
image generation.
You know, there's otherproblems because of that.
But I think people need tounderstand what it cannot do
because it's it's not magic.
Priscilla (17:29):
Yeah.
Where do you see AI saving themost time for support teams?
And do you have any examples oftasks that you've seen where AI
just does a better job than ahuman and can make things more
efficient for the human to workon other things?
Craig (17:45):
So specific to support, right?
I think the AI solutions thatare in the market, you know,
have two at a high levelfunctions, right?
FAQ, question and answer stuffthat's completely repeatable.
Everybody gets the same answer.
AI is great at that and and canbasically abolish that kind of
level of question.
Yeah.
That is sometimes high ROI.
Usually that's more medium, lowROI, because those questions
(18:09):
already were quick.
You had a macro.
I think the one advantage AIhas is that it's not macro
based, right?
So it's more generative and cansound a bit more human than a
clearly pre-written solution.
The second one is theautomation of activities.
And this is where you canintegrate in your tools and
automate things that would takea human agent five, 10 minutes,
copying and pasting, using APIs,using integration.
(18:31):
The AI can call out thesethings, recognize when they need
to call out these things, andand even checking things like
formatting, you wouldn't believehow every different system uses
a different date time format,uses a different address format.
You know, if you'reinternational, you have you have
dozens of different phoneformats that you would have to
accommodate.
And then things like language,most AI solutions are
multilingual out of the box.
(18:53):
So instead of hiring one personthat sits uh you know, sits in
Italy and can speak Italian,Spanish, French, right?
You can now an AI agent thatdoes most of that for you,
translates tickets to English sothey can be reviewed by your
Englishy if that's what you needto do.
So that's kind of the firstthing that I would say to look
at when it comes to to supportspecifically.
But to benefit the team as awhole, right, that's where you
(19:16):
know you start needing torecognize what AI is good at,
right?
And AI is great at patternrecognition, understanding
anomalous data.
And this is where it's reallycool for things like voice of
customer, or very cool for likeCSAT interactions or effort
interactions, right?
Like the customer effort scorestuff, where you could take a
(19:37):
huge data set, a month's worthof tickets, a quarter's worth of
chat, you can correlate that toyou know your customer success
export tool, your is yourcustomer in red, yellow, green
status.
You can probably nowadays wherewe have all of our record uh
calls are recorded, you couldprobably inject a bunch of that
data.
You get all this data into abig bucket and say, for every
(19:57):
customer that's in a red status,tell me the top 10 things that
they're concerned about.
Yeah, you know, and and AHI isreally good at that.
Like, really, really, reallygood at that.
That's something I thinksupport can use to heighten
their visibility within thecompany, right?
By able to say, like, go to theproduct team, go to the service
team, go to the onboardingteam, go to whatever, and say,
(20:18):
Hey, I have hard evidence thatthis is the the case, right?
And so those are those arereally kind of the places I
always tell people to start.
But I mean, there's tons ofthese things, right?
I I don't reformat anythinganymore.
Like if I have if I have datain format X and I want it to
format Y, I just upload it to anAI and say, here's an example
(20:39):
of the format, the old format,here's an example of the new
format, go.
And you know, a minute later, Idon't have to worry about, you
know, changing things from CSVto JSON or or whatever it might
be, like whatever you have to dowithin your support department.
Right.
There's tons of these littlethings.
Priscilla (20:54):
Yeah.
Well, and I think you hit onone that for me feels really
huge, which is the reportingaspect of it and the help that
that can bring.
I mean, there's so much tolearn in your support inbox,
right?
And when you're working withcustomers, there's so much data
there and insight there.
(21:15):
And when you have a supportteam that's struggling to get
through tickets, they aren'table to do all the analyzing.
And so having an AI solutionthat can come in and help do
that analyzing and turn thatinto actionable results where
you can go and take that to themarketing team or the developers
(21:36):
and say, look, everyone'sconfused about the same thing.
We've got to do something tochange this.
That can be a really huge,really beneficial way to use AI,
especially if you are a littlebit more anxious about having
the AI interact directly withyour customers.
Like that's a way to kind oflike build your trust with AI
abilities, is in that kind oflike non customer facing way.
(21:59):
Also you mentionedtranslations, which is a huge
one in the way that we use AI atBuzzsprout, is that we have
people all over the world thatare reaching out and we don't
have anyone right now who ismultilingual on our team.
And so we use the translationsolutions a lot with AI.
And it's a huge help to us tobe able to communicate with
(22:21):
people on their terms in theirlanguage without, you know,
forcing them to conform to oursjust because we're speaking
English.
So I think that's anotherreally good tool, a really good
way to use AI.
What about in the regularday-to-day, you're working in
the inbox?
What little ways do you use AI?
Maybe not even directly withcustomers, but just like to make
(22:42):
your life easier.
Craig (22:43):
I mean, anywhere where you want to sound less robotic,
like instead of using a macro,AI is great at that.
So you might have a macro, butyou could run that macro.
So run the question, you know,in in the email and the macro
and say, customize this macro toanswer the question in a better
way, for example.
I think that summarizing thingsis really good.
Priscilla (23:06):
Yes.
Right.
Craig (23:07):
I feel like when you're doing a handoff of a ticket,
like, oh, I have to go onvacation, so I'm gonna hand this
ticket off to someone that'sbeen going on for a while,
maybe.
AI can summarize and say, like,here's here's what set seems to
be outstanding.
Here's the types of things thatwe're already trying, so you're
not repeating yourself.
The tone, you know, like somepeople struggle with tone.
And and I mean, that's a that'sa skill, right?
(23:28):
To to add empathy and so run itthrough and say, Hey, like, I
really need this to soundempathetic.
We made a mistake.
I don't want to place blame, Idon't want to assume fault, but
at the same time, this messagehas to be clear.
And this is the type of stuffthat's AI is really good at.
Yeah.
You know, you could write smalllittle scripts.
I I have a few uh that I usefor troubleshooting.
(23:50):
This is a great example,actually.
So we we have some logs thatstore the entire interactions
between our customers and ourAI.
And some of the details are notexposed to the user because
they're usually too technical,or there's some backing stuff
that of course you don't want toexpose to any any end user.
Right.
But we need access to itbecause if the customer can't
figure it out, we have to dosome troubleshooting from our
(24:12):
supports perspective.
The problem is that to accessit, you have to, you know, take
some piece of information,search a database to get that
information, get the export,then look through every single
line of the database to findwhere the error occurred.
And that was a time consumingprocess.
I wrote a small tool.
And by wrote, I mean I vibecoded using, you know, using
cursor, because I'm not strongat Python at all, but I wrote a
(24:34):
small script that I would giveit uh the identifier, whether
that be a keyword, what we calla conversation ID, which is a
unique identifier for everyconversation.
It goes to the database, getseverything related to that
conversation.
And then not only does it giveyou the export to a Google Sheet
of all the raw data so you canreview it.
I wrote a prompt that wouldtake all that data and spit out
(24:57):
a report on that data.
So, what did the AI find?
You know, talk about anomaliesand pattern recognition.
I would say, like, did you seeany failures in the
integrations?
You know, that's a big cause ofproblems, right?
Um, did you see any repetition?
Did you see all these things,like a list of 10 things to
identify?
And it spits out a report.
And now you could start thetroubleshooting instantly.
(25:18):
It takes about uh 30 seconds torun all that through the
script, and you have a reportthat in theory you could just
send to the customer and say,Well, here's the root cause of
the problem.
That is super powerful, right?
If you're dealing withtechnical support.
And AI is awesome at that.
Priscilla (25:33):
Yeah.
Craig (25:33):
Because it it's all pattern recognition, and that's
what AIs are good at.
Priscilla (25:37):
Yeah.
Another area that we use AI ishelp us with research.
There's a lot of support thatwe get questions about adjacent
software or platforms thatpeople are using, where
typically you would expect thecustomer support team to say,
hey, sorry, we actually don't dosupport for X over there.
(25:59):
We do support for Bud Sprout.
And so we can help you withthis, but I'm not an expert on
X.
And I'm just saying X as arandom example, just realized
that X is also a platform.
So didn't do that on purpose.
Um, but it's it's one of thosethings where you kind of feel
like, oh, I'm kind of trappedhere because I don't know how to
answer the question about thisother platform.
And that's what they're askingme about.
(26:20):
And a way that you can use AIto level up that response is to
do a little bit of researchthrough AI and say, okay, here's
the question they're asking.
Can you lead me to the righthelp article or can you explain
something to me?
And then it allows you as thesupport specialist to get a
higher level of knowledge and todeliver that to your customer,
(26:41):
which is so big when you're notworking for that company.
So you're now giving them acomplete experience with the
help of AI.
Craig (26:51):
I love that.
In fact, I actually just didthat for one of our customers.
They had a CSS problemembedding our widget on a
specific platform that theywanted to embed it on.
And I was like, I've never seenthis platform before.
I don't know anything about it.
And I did exactly that.
Yeah.
Exactly that.
It just went and said, okay,well, I can't answer the
question, but I can tell youthis is where we need to look
somewhere in this area.
And right.
Yeah.
(27:11):
Great example.
Priscilla (27:12):
Yeah.
And kind of on that same laneis brainstorming, which is
something that I really love.
I think it's probably the way Iuse AI the most in my day to
day is to brainstorm things.
I'm a very verbal processor.
So what I would love to do issit down with another person in
a room and talk throughsomething.
That's not always going to bean efficient way to solve a
(27:32):
problem.
And so being able to sit withAI and throw out ideas and get
feedback and kind of brainstormthings.
What am I missing and help fillin those gaps?
I think that's another way touse AI, especially when you're
in the support inbox and you'retroubleshooting an issue.
Say, okay, I'm here where weare, here's where we've tried.
What am I missing?
Like, what am I not seeing?
(27:54):
Those kinds of things can be areally helpful way to just use
AI in your regular, you know,daily process.
Craig (27:59):
I completely agree.
If I start so many of my AIprompts with, I'll have my
headset in and I'll be walkingsomewhere and I'll be like, I'm
in a rant for a while, and thenyou're gonna summarize what I
say into something clean and youknow, pull out what you feel
the most coherent thoughts areand most relevant.
When I'm alone in the car, I'llI'll use to write emails.
I'll just plug it in and say,like, I need to write three
(28:20):
emails.
One is about this topic, andhere's what I want to generally
say.
Next, here's, you know, and Ioften don't send them verbatim
because I find that even tone ofvoice still is a bit off, even
if you've tried to train it.
Yeah.
But at least it gives me theframework and I can just copy
paste and then change the fewwords, which I still think is is
faster.
And I can be productive, youknow, in in the car sometimes.
But you know, formattingdocuments, you know, like a root
(28:42):
cause analysis is sometimes abig one in support, depending on
your sport team, right?
You might have to deliver aroot cause analysis doc.
Don't make up a format.
Don't try to figure out thebest way to highlight things or
what's the what should go in theexecutive summary.
Upload all the data about theproblem, whatever it was that
you're root cause analyzing, andsay, build me a root cause
analysis report, and it willspit you out something that is
(29:05):
in a format.
Again, you're not relying onknowledge there.
You're relying on the fact thatit's betrayed on thousands and
millions of the root causeanalysis reports that it can
give you the formatting thatmakes sense for your use case.
And yeah, all of those usecases are ways that you can just
save minutes or hours.
And you know, going back towhat you said, people are you
(29:26):
know fearful of AI and someone'sjob is not going to disappear
because you are reusing the bestpractices on how to build a
report, right?
Priscilla (29:35):
Right.
Well, and I think that that'swhy conversations like this are
so important because the peoplewho are fearful of losing their
jobs, the people who are worriedabout the stereotype of, okay,
this AI chatbot is going toreplace me, and that is the
extent of AI and customersupport.
Hearing these conversationswhere we're breaking it down in
(29:56):
a way that's like, hey,actually, this is something
beneficial to you.
It's important to hear thesekind of conversations where it
takes it out of this like scaryunknown to, oh, okay, you're
talking about using AI as a wayto brainstorm, or you're talking
about using AI as a way to helpyou analyze your tickets.
Like those kind of things, Ithink, really help and take away
that fear factor of thischatbot is gonna do my job
(30:19):
better than I can, and now I'mno longer gonna be necessary.
Okay.
So when I think of my earliestconversations about AI here at
Buzzsprout, one of the thingsthat I was worried about was
that we had built up for thelast couple of decades this
relationship with our customersand this expectation where they
(30:40):
knew that if they emailed us,they were gonna get a response
from a human really quickly.
And so I think it's reallyimportant for us to talk about
some of the mistakes that peoplemake when they implement AI
because there is a lot of riskthere because you are working
with your customers who are, atthe end of the day, the most
important part of your business.
So, you know, what have youseen kind of as examples of what
(31:01):
not to do when people areimplementing AI in maybe not the
best way?
Craig (31:06):
Sometimes the tendency is to go back to uh decision
tree-based logic.
That's the first mistake that Itry to talk to our customers
about is the benefit of AI isyou don't have to do decision
trees anymore.
For example, if you're aclothing company and you want
to, you know, help people pickthe right size of the clothing
(31:27):
you sell, there is a set ofthings that you need to know
about the person's purchase.
If they're a man or a woman ora child.
If it's for a shirt for a man,you probably need chest size.
If it's for a uh blouse for awoman, you might need uh bust,
waist, and hips.
You have a set of things youneed to know.
And the historic way of solvingthat was to have a decision
tree.
Are you a man or a woman?
(31:48):
Man, oh okay.
Are you buying pants, shirt,you know, pants?
AI, you could just teach itthat.
Here is the things you need toknow.
Here's where that informationis.
Whatever you don't know, figureout with the customer or infer
based on the customer's profilethat you have access to or the
context that you have access to,and then only ask the questions
(32:09):
that require.
Don't repeat them, stone havethem repeat themselves.
You have context of who thatperson is, put all that together
and only ask the questions thatyou might not have an answer
to.
Using macros is a reallyinteresting one.
We have customers who have thisbeautiful generative model that
can generate reallyhuman-sounding, personalized
messages, and they want to reusethe macro text from their help
(32:30):
desk and just set use our AI tosend the same text as the the
help center would have.
And the last one, and probablyone of the biggest mistakes I
see is people, and this is athis is one that gets to me more
of a personal level, and thisis maybe my hot take for the the
podcast.
I love it.
CSAT for AI is different thanCSAT for a human.
And so many people are solocked in, dude.
(32:55):
We have to have CSAT.
We have to have CSAT.
And they forget that CSAT is isnot really what you need to
measure in AI.
Data shows, you know, if you ifyou were to look at all of our
customers and export the CSATdata, it's always slightly below
average.
And that's what you wouldexpect for a neutral, like
people use AI differently andreport on it differently than
(33:18):
they would with a human.
Because if I'm asking, do youship to Germany?
And the answer is no, that's anA and a pretty neutral
question.
And if the answer, it's eithergonna be, yeah, oh, you know,
I'm sad because, you know, it'snegative because you said no, or
it's positive because you saidyes.
Like there's no, there's nonuance to CSAT like there would
be with a human, where it'slike, oh, their tone was they
just treated me poorly, and youknow, I'm gonna give them a
(33:40):
lower score, right?
What you really want to measurewith AI is the things that
frustrate people about AI.
Did they go into a loop?
Did they ask the customer torepeat themselves?
Did the AI misunderstand in thecustomer and say, no, that's
not what I meant?
Right.
Those are things you should bemeasuring.
I call that frustration index,and that's something that our
product measures because CSAT tome is not as relevant with AI
(34:04):
as people think it is, becauseit's just a totally different
concept when it comes to howpeople aren't frustrated by
humans differently than they areby bots because they treat bots
differently than they treathumans.
Like it's it's not an apples toapples comparison.
So that's one that I alwayslooked at.
And coincidentally to yourquestion, is that everything
that I put in that list offrustrated index, I would label
(34:26):
as the mistakes you do when youimplement AI, right?
If you allow your AI to getinto loops, if you don't train
it on synonyms well enough sothat it understands the context
and and and misunderstands yourusers, you know, if you don't
have the right guardrails inyour product that prevent
hallucinations and sell you uh,you know, a one-dollar car, like
that Chevy example or the Idon't know if the Air Canada
(34:46):
example where they gave away abereavement flight that didn't
exist, you know, made it.
But that's that's the problemwith AI.
That's bad AI.
Yeah, right.
It's that those are themistakes you make, is that you
don't put those guardrails inplace.
So those are kind of the maybethe top three or four examples
that I have of people like, youknow, implementing AI poorly.
Yeah.
And that you need to avoid.
Priscilla (35:05):
And I'll add my own to that, which is not giving
people a way to get to a human.
Craig (35:11):
100%.
Priscilla (35:12):
I think that a lot of times I will hear these stories
of people like jumping into AI.
It's so exciting, it's sogreat, but they don't give you a
route out to a human.
And I think that's a hugemisstep to not have that,
especially if your users areused to working with humans.
And then now you put this AI infront of that, you're gonna
lose trust if they don't have away to get to you.
(35:34):
And I think you said itearlier, this idea that, you
know, it's not that they don'twant to work with AI, they just
don't want to work with bad AI.
They don't want to work with abad chatbot that's gonna throw
them in the loop.
And so you're gonna run intothose situations.
And I think that frustrationindex is a really good way to
measure the success of your AIis okay, how much are my
customers getting stuck in thesefrustration cycles?
(35:57):
Because that's what you want toavoid when you're implementing
AI.
You wanna make sure you'regiving a smooth, top-notch
experience because I thinkcustomers are gonna have a
shorter fuse when it comes toworking with AI than with a
human when you're like, givethem a bad response.
Immediately it's gonna be like,nope, I knew I knew this was
gonna be bad and now I want out.
You know, so you have to bereally careful with that.
(36:20):
Do you have any funny storieswhere AI just got it wrong?
Craig (36:23):
I would say it's wrong, but I have a couple of funny
ones where we had a particularclothing brand and it was very
flowery, very summary, like thatwas their brand.
And they speak with the poetic,you know, nature of Maya
Angelou was something again,very kind of like poetic and
summary and happy.
And so when we were buildingthe tone of voice, um, we put
that in there, you know, speakwith the poetic, you know,
(36:46):
angles of of Maya Angelou.
The bot was hilarious.
There's some of my favoriteones that came out of this was,
you know, it was things like,you know, what's your shipping
policy?
It's like to traverse the worldwith through the winds of
change, as the, you know, and itwas just it went on and on and
on.
It was like, we will take it toEurope, we will take it to
Asia, and we will even cross thePacific over into the Americas.
(37:10):
And like it was just that wastheir shipping policy as
interpreted by that.
And then we had anothercustomer, they were uh a company
owned by a famous uh rapper,and so they wanted to have that
kind of built into the samething in the tone of voice, and
and the same thing.
The the the questions came backwith this like real beat, like
you know, kind of nonsensical,lots of slang, lots of so those
(37:32):
are a couple of funny things,but I can't think of an example
where an AI has gotten itcompletely wrong.
I think there's some reallyinteresting things that you
don't think about when you'retraining AI.
So, for example, we this wasactually a Kodif example.
So we use Kodif on our websiteas you might expect, and someone
was asking for a phone numberto contact us, and we don't
really have a receptionist, likewe don't have a phone in our
(37:53):
office.
We're a small startup, but theAI, in one of our privacy policy
documents, found a phone numberfor an old office when the
company was first incorporated.
So the document was out ofdate, but but it found this
phone number, and it's like, oh,the only number I have access
to is this phone number, whichthe person then called, and of
course it was disconnected.
And so it was one of thosethings where it's like, well,
(38:14):
where did it find the number?
But it was just so funnybecause someone was like, Well,
how did it, where would it havepicked up a number?
And it did it hallucinate thisphone number, but it didn't
hallucinate it.
It was it was in the trainingthat, oh, it had said if you
looked at the document tocontact Kodif for uh a privacy
or security concern.
Here's this phone number.
Yeah, just stuff like that,where we get these customers who
(38:36):
who ask us, well, how did itcome up with this response?
And you look and it's like,yeah, you know what?
It actually it makes sense itcame up with this response
because of these uh theconnections it's making here,
even though those connectionswouldn't wouldn't have been made
by a human anyway.
I love the I love the TotalVoice things because I think
that those are reallyinteresting interpretations.
Priscilla (38:56):
Well, and I think it speaks to the importance of, and
you were talking about this alittle bit ago, but the idea
that like our skills are goingto change now for customer
support specialists, becauseyou're gonna need to lean into
these other skills in order tocorrectly train the AI and those
strategies.
And so it kind of leads wellinto my next question, which is
(39:18):
about those skills, because youneed to be able to have people
on your team who can train theAI in the right tone to navigate
those kinds of um examples thatyou just gave, where someone
asked for a certain kind of toneand it didn't react in the way
that is gonna ultimately be theway you go moving forward.
So, how do you train people ontraining the AI?
(39:38):
So, what other kind of skillsdo you see needing to be
developed?
Say you're in customer supportand you're looking at the future
of your career and you're like,I want to invest in some of
those skills now.
What are those skills thatpeople are gonna need to have
sharpened going forward incustomer support?
Craig (39:56):
So, I mean, that the number one is just understanding
how AI prompts work.
Priscilla (40:01):
Yeah.
Craig (40:01):
Right?
AI prompting is a is going tobecome a skill.
And in fact, I would argue it'ssomething that they should
start teaching in in the highschool areas.
Yeah.
It's gonna become thatfundamental.
And again, in five years,probably I'll look back on that
statement and say how stupid Iwas, because now AI can
interpret anything.
But yeah, right now, AIprompting is so vital.
And what I mean by that is thatlet's go back to the root cause
(40:22):
analysis report example.
In a world where you'reproducing that report, you know,
the skill is to logically takefactual data and analysis and
put it into some logical orderto read, right?
That's the skill.
In a world where you're notproducing the report, you are
sharing the information withinwith an LLM and asking it to
produce the report based on theinformation.
(40:43):
The skill is to describe theinformation, what are the key
factors it has to highlight,what are the things to ignore,
what to point out, what not topoint out.
And there's a specificity tothat, right?
There's a skill to being ableto be very precise in your
language to tell something howto do something.
And uh the better you're at atdoing those things, the more the
(41:06):
report will sound like you, themore the report will look like
how you want it, the lessediting you'll have to do after
the fact.
But that bit is going to becomethe skill.
And I think the number twoskill is we have to get into a
mindset of this personalizationmore as humans, too.
I had a really cool experiencewith Air Canada recently where I
had to change something on aflight and I called, and the
(41:28):
first thing it said to me beforeit even started ringing was
based on your phone number, uh,we think you're calling about uh
your flight to Denver nextweek.
If this is correct, press one.
Wow.
Right?
Like, wow, seriously, becauselike now, how many times have
you tried to say those six-digitalpha numeric codes in a broken
NATO alphabet?
And N, you mean N as in Nancy?
(41:50):
No, M, M, N, you know, yeah.
That was a beautiful use of notreally even AI, it's just use
of technology to get that andpersonalization.
Personalization, but becausethat's what people are going to
get used to with AI, humans haveto adopt that mentality now and
be like, oh, if I have thatinformation available, if I if I
see a phone number with a name,or if I see uh the person has
(42:13):
you know purchased this oneproduct recently, again, don't
ask them to repeat themselves.
Don't ask them to spelleverything out for you.
Use the context to help them.
Start the conversation with uh,well, how did you like that
lipstick that you bought lastweek?
You know, like there's a way toadapt that into your style and
tone of voice.
And I think again, that's askill you're gonna have to teach
(42:33):
uh your team.
And I don't think it'sunreasonable to say that that
will become an expectationbecause if a bot can do it, a
human should be able to do it.
Priscilla (42:41):
Yeah.
Well, I think it's one of thecoolest things about AI and
customer support, is it's goingto give customer support
specialists the ability to geteven more personal and to
personalize things and to focuson building relationships
instead of just closing tickets.
And I think that's really whereif you're incorporating AI into
your workflow so that you canfocus on creating stronger
(43:05):
relationships with yourcustomers, you're doing that in
the right way.
Right.
You're you're looking at itfrom the right perspective.
If you're incorporating AI toeliminate a support team, that's
probably not the right route totake.
In five, 10 years from now,when we're using AI still, what
aspects of this customer supportrelationship environment, like
(43:28):
what are the things that need tostay human?
Craig (43:30):
Things like empathy and situational re uh judgments and
reactions are always gonna haveto be in a human form, right?
Yeah.
It's so hard when you say five,10 years, uh, because I just
this this technology is movingso fast.
So fast.
It's one of these things that Istruggle to know what my seven
year old is gonna do in theworkspace, right?
(43:51):
Not because I don't think he'sgonna have a job, by the way,
just as we already talked about.
But so when it comes to whatshould stay home, then I think
it's a stuff that is uniquelyhuman.
Situational reasoning, I think,absolutely is there.
There's there's no, you know,this artificial general
intelligence, the AGI that thepeople talk about.
I'm my God is it's not five, 10years away.
Like we're it it just doesn'tseem possible to me.
(44:13):
I don't think it's a hundredyears away, but I, you know,
creativity, like solving aproblem.
You you mentioned, you know,you do research with with AI,
which again, I do too.
I think it's a smart idea, butthat's not ever going to
generate a new idea.
It's gonna validate otherideas, it's gonna say, hey,
other people have done it thisway.
It's gonna help you absolutely.
Yeah, no criticism, but it'snot gonna generate a net new
(44:34):
idea.
It can't, because that's nothow AI currently works.
Right.
AGI would have to exist forthat.
And so that's the stuff.
Anything, you know, being ableto, again, from a support
perspective, build a creativesolution, like, oh, I'm sorry,
we're not gonna be able toreship this thing because it's
out of stock, but here's what Ican do because you know, I
you're a valued customer, right?
(44:55):
Like that type of creativity,uh, of problem solving.
And I think anything to do withreally deep product or service
or whatever industry knowledge,like that deep level where it's
like, I understand exactly whatyou're trying to achieve with
our products and services, andI, as an expert in that area,
(45:16):
can help guide you down a pathto success for this specific use
case.
Again, that that's justsomething that if you, you know,
you might be able to train anAI on on some of that, but I
don't think it's gonna have thatdeep, full understanding that a
human conversation can have.
And so, yeah, those are areasthat I think are uniquely human
that that probably won't go awayuntil AGI exists.
(45:38):
And even then, AGI is not magiceither.
Yeah.
It still has to have this basisof how to operate.
And I do picture a world whereI call it the big white box.
That's my my name for it.
Everyone is gonna have this bigwhite box, whether it be a
website or or maybe it's aphysical box.
It doesn't really matter.
And that thing, that box knowseverything about you.
(45:59):
It knows your age, your height,your clothes size, it knows
what type of dishwasher youhave, it knows where you live,
it knows that your floors arecarpet, it knows whatever.
It knows everything.
And when you have an issue,you're gonna go to that big
white box and describe yourissue, and it's just gonna
figure out how to do it, whetherthat be contact a human or a
sport department, whether it beto call a service person to come
(46:22):
and fix it, whether it be tobuy something net new, whatever.
The box will know everythingabout you, will know your
spending habits, budgets, itwill know everything.
And you just say, you know,hey, my upstairs tap is leaking,
and it's gonna be like, oh,well, that tap is still under
warranty from Mowen.
I'll submit the warrantyrequest, and here's the place to
do it, and I know your address.
(46:43):
So I'm you don't even have tofill it in.
It'll be there tomorrow.
That's coming, right?
Yeah, so that's a differenttype of support, right?
That is someone, you know, fromevery all these companies
plugging into the big white boxto say, you know, if I want to
be a plumber and I want aservice call to these houses,
that big white box needs to knowabout my plumbing services and
my service radius and my pricesand my and so there's another
skill.
(47:03):
You know what I mean?
Like it you're gonna have toteach this big white, and that's
but that's where we're going,right?
And and so the you know, goingback to what the human element
is of that, it's the teaching ofthe big white box, how to get a
hold, how to start a warrantyrequest for a dishwasher, how to
hire a carpet cleaning servicebecause you spilled wine on it.
Like that's the thing.
That's the next you know,another layer of how support's
(47:25):
gonna get there.
Priscilla (47:25):
Yeah.
I feel like we've touched on somany things in the last you
know, hour.
I'm gonna ask you one lastquestion to wrap this up.
What is one truth about AI thatyou want people to take away
from this?
Craig (47:37):
It would be disingenuous of me, especially someone who's
in the industry, that it's notscary, right?
There's a lot of unknowns,there's a lot of things that
could go wrong with it.
You know, there were peopleback in the 90s.
I remember people saying thisinternet's gonna destroy
civilizations, and and to somedegree, maybe it has.
Maybe it has.
We can we can have thatargument as well.
(47:59):
But the thing is, is that it'sit's all about adaptation.
We should be trepidatious.
I just don't think we should beuh uh A, completely ignoring
it, and B worried that it'sgonna cause mass global
unemployment.
Uh that's just not so the truthto me is that there's a middle
ground where we accept that thisis a technology that now
(48:20):
exists, it will never not exist.
And so, how do you want to useit in your daily life?
And everyone can choose how toenable it or how not to enable
it.
And that's that's really got tobe up to up to you.
And then the rest is up to ourleaders to to help society
thrive with it.
So, yeah, that's probably my myopinion on the truth of AI.
Priscilla (48:39):
Well said.
Thank you so much for beinghere, for sharing with us,
Craig.
This was really informative.
I hope that our listeners areable to take some of this into
their daily work.
If they want to learn moreabout you, and if they want to
learn more about Kodif, how canthey learn more about you and
Kodif and how they could use it?
Craig (48:57):
Well, I'm LinkedIn.
That's the main place I am.
So look, I'm the only CraigStoss on LinkedIn.
I'm almost guaranteeing you ofthat.
Nice.
Um and uh and then Kodif is isk-odif.io.
But yeah, I mean, please reachout.
I love having theseconversations.
So if you have questions, if ifyou want to discuss something I
said or you know, if youdisagree, I I love hearing
(49:19):
disagreements too, becausethat's how we we build better,
uh, better content for people.
So yeah, that's uh that's howyou get a hold of me.
Priscilla (49:26):
Yeah, and it's good to have these conversations
about AI and to learn more aboutit.
And that's the way thingsbecome less scary, is when you
understand them more.
So if you are listening to thisand you're scared about AI,
reach out to Craig, have theconversation with him, and
you'll learn more and it'll takesome of that fear away a little
bit.
If you have a question or atopic that you would like us to
cover on a future episode, youcan shoot us a text using the
(49:49):
Send Us a Text link in thisepisode description.
As always, if you like theepisode, please share it with
someone who works in customersupport or leave us a review on
Apple Podcasts.
Thank you again to Craig Stossfor joining us on today's
episode.
Thank you for listening.
Now go and make someone's day.
Welcome to Happy to Help, a podcast about customer
support from the people ofBuzzsprout.
I'm your host, PriscillaBrooke.
Today we're getting practicalabout AI and customer support.
We'll talk about commonmisconceptions about AI,
actionable tips forincorporating AI into your daily
workflow, and what you need toknow to get the most out of your
AI strategies.
(00:20):
Thanks for joining us.
Let's get into it.
All right, so today is one ofthose episodes where I'm really
excited.
I know that at the end of thisepisode, everyone listening is
going to have some really goodtangible things to take and put
into their workflow tomorrow.
And so those kind of episodesalways get me really excited.
(00:42):
I feel like we've all listenedto podcasts where they give you
these big overarching ideas andconcepts to think about, which
can be super helpful.
But sometimes you just wantsomeone to tell you what to do
and how to handle something.
And so I'm really excited thatwe're going to talk about those
practical uses of AI, especiallyfor the people listening to
(01:03):
this who don't use AI much orare kind of scared of it or have
a smaller team and don't, youknow, have the funds or the
budget to buy a big AI platform,like how they can use AI
strategies in their day-to-day.
So joining us today to talkabout AI is Craig Stoss.
Craig is the VP of Solutions atKodif, which is an AI-driven
(01:24):
customer experience and supportautomation platform for
e-commerce customers.
He has over 20 years ofexperience in customer-facing
roles and is passionate aboutthe impact that customer
experience can have on yourbusiness.
Today, he's going to join usand bring us his insights into
AI and give us those practicaltips about using AI and customer
(01:45):
support.
So thanks for being here,Craig.
Craig (01:47):
Thank you for having me.
I'm so I'm so excited for thisconversation.
I think it's so relevant rightnow to so many different types
of businesses.
Priscilla (01:53):
Oh, yeah, for sure.
And I feel like it's one ofthose conversations that gets a
lot of people, whether they'resuper excited about AI or
they're super hesitant.
Like it's just there's a lot offeelings and emotions that come
along with AI.
And so it's always a goodconversation to have to like
break it down and look at it ina way that can kind of take that
emotion out of it.
We always kick off our episodeswith the same question.
(02:16):
So, Craig, who has made yourday recently?
Craig (02:19):
So my son and I went to get our haircut.
And the barber we have has akind of electronic waiting board
where you check in, you put youput your your name in.
But the way it works is whenthere's multiple people in one
group, you're supposed to selectthem all at once and hit go.
So you're kind of groupedtogether.
And I just forgot.
And then, of course, before Icould get my son checked in,
(02:41):
another person had checked in.
So now there was this big gapbetween me and my seven-year-old
son.
And I thought, oh great.
I'm gonna, you know, get myhe's gonna get his care haircut,
and then I'm going to have towait for 20 minutes and again
get my hair cut.
And he's gonna be sitting inthe, you know, getting bored.
And it was already about a 30,40 minute wait to get uh seen,
anyways.
(03:01):
Wow, yeah, it was a very busySunday at that the barber shop.
And I was I was reallyregretting this because when my
son gets bored, it's chaos.
And oh, and just to end theboot, they had a TV in the
lounge that the uh the TV wasnot working that particular day
on Sunday.
So that even made it worse.
He couldn't even be distractedby the TV.
Jordan (03:18):
That's unbearable.
Yeah.
Craig (03:20):
Yeah, yeah.
So I chatted with the woman whowas doing my hair and she was
very nice.
And and without a without meeven asking, she accommodated
and said, listen, we'll justreassign you.
And I thought, you know what?
That's a recognition of the thesituation that that you're in,
right?
And and the other person didn'teven notice, right?
You know, because my son's hairtakes two minutes to cut
(03:40):
because it's he's small.
So that made my day in a sensethat it was a really small thing
that, you know, bending therules a bit, but uh, but it
really it really helped probablyeveryone in the situation so
that my son didn't have to runaround the the barbershop going
crazy for 20 minutes.
Priscilla (03:57):
Yeah, it had it benefits and impacts that people
didn't even realize, right?
Craig (04:01):
Exactly.
Priscilla (04:02):
It's funny, I was talking about a similar
situation just this morning.
We were talking about refunds,and I was thinking about how
sometimes working in a situationin customer support, and it's
clear that a refund is the rightthing to do, whether they've
asked for it or not.
In the situation I was talkingabout this morning, someone had
upgraded to a paid plan, butthen immediately decided that it
(04:23):
wasn't gonna work for them andthis downgrade.
But they didn't ask for arefund and the specialist
offered one anyway.
It was one of those thingswhere the impact of the offer,
because it hadn't beenrequested, was even more than it
would have been had the personasked for it.
So if you had said, hey, canyou switch this around for us,
they probably would have done itand it would have been fine.
(04:44):
But the impact of the personnoticing that you needed that
without you having to ask forit, that's such a bigger impact.
And so I think it's one ofthose, like, as a customer
support specialist, being ableto identify when the right thing
to do something, even if it'snot requested, can be a cool way
to handle a situation.
And it shows like you careabout the situation more than
(05:07):
just reacting to the request.
I think that's a great story.
And shout out to them formaking that easier for everyone
involved.
Before we kind of get into AI,can you tell us ju"st a little
bit about how you got intocustomer support, but then also
a little bit about the workyou're doing now, kind of at
that intersection of AI andcustomer support?
Craig (05:26):
You know, I grew up in the 80s and and you know,
computers and things were justnot really a thing.
The internet hadn't beeninvented yet.
And so what I finally did getto touch my first computer, you
know, it at the time it wasn'tseen as it is today, where it's
like, you know, the internet andgames and flashiness, right?
It was it was a tool.
It was, it was, you know, justa fancy calculator.
(05:47):
And so, you know, I spent thefirst year coding and being so
fascinated that I could talk toan inanimate object and have it
do things that I asked it to do.
And as I kind of developed anddeveloped and developed that
kind of skill, and I got intoum, you know, there was a
program in Canada and ruralcommunities to bring when the
internet didn't go mainstream inthe 90s, there was a program in
(06:09):
in rural communities Canadawhere they would hire students
to teach senior citizens aboutwhat the internet was and how to
get an email address and how toget government forums.
And and I was selected to runthis kind of program.
And that was right there that Ireally saw this intersection
between technology and the humanparts.
And and from there I really didback down.
(06:29):
I did take a computer sciencedegree.
Uh, it was mainly in like youknow, the full life cycle of
software development.
It wasn't just development.
I learned how to writerequirements documents, I
learned how to QA, uh, I learnedhow to present ideas and gather
information from a customer.
All my internships during that,the during my university years
were in support departments orat least in customer-facing
(06:52):
departments.
And I just really felt likethat was where I wanted to be.
I didn't want to developfull-time, I wanted to kind of
be that intersection betweentechnology and people.
And over the years, I did I'vedone lots of different roles.
I was consultant for Fortune500 companies for software
development best practices.
I got into support leadershiplater after that.
Most recently running, uh, Iwas at a BPO and then uh running
(07:16):
some of their customer programsuh for their teams.
And then now in the AI world.
And the thing that's alwaysfascinated me about software in
general is support departmentshave all have this kind of like
cost center mentality.
And so software and softwaresolutions for support have
always been this kind ofcommodity in some ways, right?
(07:37):
It was one of these things thatthere's tons of cool solutions,
there's tons of systems.
10 years ago, when I first gotinto leadership, I was pitched
on tagging solutions andcategorization solutions and
reporting solutions andchatbots, early versions of
chatbots.
Yeah.
That stuff was just stuff thatyou would have a hard time
affording.
Yeah.
And what I where I think AI isdifferent, and what's really
(07:57):
cool about the AI industry andwhere we're going is that these
things are now more affordable.
The ROI, given the scale andscope of most AI solutions, is
much more clearer than say alegacy knowledge management
system.
You know, it it's very easy tosee that.
And so, yeah, that's how I gotinto it.
(08:19):
And and now that's how itintersects into my current role,
where I'm still in thattechnology and I'm still, you
know, as a solutions person, I'mstill talking to customers
every day about how they can usetechnology.
And that's ultimately mypassion is about making these
experiences high quality forthem.
Priscilla (08:34):
Yeah, I love that.
Can you tell us a little bitabout Kodif specifically and
what Kodif does and what itoffers for its customers?
Craig (08:40):
Yeah, I've known of CODA for about five years.
I met the founders at aconference and we got talking
about the future of AI, and itwas just on that cusp of Chat
GPT and open AI solutions comingout.
Priscilla (08:53):
Yeah.
Craig (08:53):
And one of the things that I've always believed in,
and this has nothing to do withAI at all, is that support needs
to be personalized andcontextual, always, right?
And CODA, when I was talking tothe founders, was one of the
first AI solution companies Italked to, and I've talked to
dozens at this point, thatreally was how it was built from
the ground up was everythingthey did, all their prompts, all
(09:15):
their back-end work was aboutmaking support contextual and
personalized.
So if you say something like,My grandfather lives in Germany
and I want to ship them yourproduct, it's not just gonna
say, Oh, I'm really sorry wedon't shift to Germany.
It's gonna say something like,Oh, what a great idea that you
want to ship something to yourgrandfather.
I hope he has a wonderfulbirthday, but I'm really sorry
(09:35):
we don't shift to Germany.
Wow.
You know, that's a much morehuman-sounding message.
It takes in the context yourquestion, it personalizes it,
those types of things.
And so every feature we buildhas like really strong abilities
to categorize, to understand,to empathize.
Like we have a lot of pet foodcompanies, and sometimes when
you're a pet food subscriptioncompany, people have to cancel
(09:57):
their subscription because a pethas died.
Priscilla (09:59):
Right.
Craig (10:00):
And you want to opt that empathy and say, Oh, I'm really
like that's not the time whereyou go in and be like, oh, well,
we'll offer you 25% discount ifwe if you were staying with us.
It's like, yeah, no, theperson's just lost a family
member, right?
So but that's what really drewme to Kodif initially, and and
that's something that goes intoevery feature we we produce.
Priscilla (10:18):
Wow.
Obviously, you're about to getinto all of this as far as you
know, how to use AI in a waythat is personal, is human, and
includes that empathy and stuff.
So let's jump into kind of likethe early days of AI and
customer support.
I remember when AI first waskind of coming on that it was
(10:40):
really scary, right?
And it was like, okay, is thisjust gonna like take away all
customer support roles?
And I had just really startedto solidify my existence in
customer support and just lovingit as work.
And then, oh no, AI is comingin.
Is this just gonna take it allout?
We've been working so hard tobuild this up.
And now are we gonna just giveit over to robots?
(11:00):
So I'm curious, you know,looking at how it's evolved over
the last three to four years,when you started working with
customer support teams, Craig,to implement these AI
strategies, what were some ofthe early reactions you were
hearing from customers and howhas that kind of changed over
the last couple of years?
Craig (11:19):
First and foremost, everyone's associated AI with
chatbot.
Those two terms were synonymousin many people's heads.
And I think the first thing wealways had to dispel was that's
not true.
I think that the most commoninterface for a AI tends to be a
chatbot, but that's not,they're not the same.
And the GPTs and the claws andthe uh and the Geminis of the
(11:43):
world are not anything relatedto the historic chatbots that
we're used to.
And and so that was the veryfirst thing was kind of getting
into this mode of why is thisdifferent or why is this better?
Because initial chatbotsrequired you to train it on kind
of questions and then answers.
(12:03):
So here's a list of questions,here's a list of answers, right?
Here's how they pair together.
And for the most part, therewas no kind of context of
synonyms, there was no contextof of different ways of asking
questions.
It was it was really, you know,match this keyword and then
this thing might be thisquestion, and here's some
answers.
And and you you remember thisfrom like early Zendesk AI days,
(12:25):
where like it would send you aresponse, and it was like these
three articles might help you.
And they were sometimes justway off base.
They just happen to mention thekeyword you used, right?
And that's not a shot at ZenDesk, right?
And I mean, that's thetechnology that we had.
Yeah.
But when when you talk tocustomers and that's what
they're associating with AI, youhave this upward battle of
saying, no, that's not whatwe're talking about.
(12:47):
We're talking about, you know,large language models, we're
talking about contextuallanguage models that understand
synonyms, they understand uhlanguage, they understand
colloquialisms, they understandthese things.
And I use the word understand abit loosely.
I think there's there'sprobably someone screaming at
their phone right now, theydon't understand anything.
And that's true.
Understand is a loose termhere, but they are trained on
(13:08):
this and they they under theyunderstand the connections, I
guess is the better way ofphrasing it.
Priscilla (13:12):
They can navigate it, yeah.
Craig (13:13):
Yeah.
So that's the first of theearly reactions.
I think how this has evolved isnow that in user-focused
industries, people have seen alittle bit about how AI is
improved.
They see these littlepersonalizations, they see that
the answers are not guessesanymore.
They're they're fairlyaccurate.
And so there's there's a littlebit more of that acceptance.
(13:35):
And when it's done well, andthere, you know, being able to
say something like, I would liketo speak to a human is a
seamless process to get to aperson or get to a human for a
certain type of ticket, or do sowithout you even, you know,
going back to the hairdresser,doing so without even asking,
just recognizing, oh, thisperson's frustrated.
Let's get them to a human.
Right.
That technology enables that.
And I think that that peopleare seeing that this is now a
(13:57):
benefit uh overall.
So there's definitely beenacceptance very recently.
Priscilla (14:02):
Do you see any current misunderstandings about
AI?
Is there like this far into it?
Are there things that you seekind of trends where people
misunderstand things about AIand how it comes into customer
support?
Craig (14:15):
Yeah, there are two big things that I would I would
argue uh I encounter on aregular basis.
The first is AI cannot makejudgment.
I think that's a reallyimportant distinction.
So, what we just talked aboutwith rules, AI has to have the
rules in place.
It can make judgments to apoint where you can teach it how
(14:39):
to make a judgment, but at thatpoint, it's not so much a
judgment as a different type ofrule, right?
And and so you know, I hearcomplaints all the time.
I was actually, we're going ona trip with some other couples
in a in a couple weeks, and wewere talking about because we
all have children, and they weresaying, well, every airline
seemed to have a differentalgorithm to assign seats with
adults next to children, andwhich is true.
(15:00):
It's interesting because Iwould argue that a lot of that's
done using some version of AI.
There's some algorithm inthere.
It may not be traditional LLMAI as we're talking about, but
there's some algorithm in there.
And it's like, well, they'reclearly just not putting,
they're either purposely notputting the rules in correctly
to make you buy a seat and forceyou to, or the rules are wrong,
(15:21):
right?
Because that's something Ithink you can define.
So going back to the judgmentpiece, so if I say to it, here's
how to judge if someonedeserves a refund, and then say
things like, you know, if theorder is greater than 30 days
old, they cannot get a refund.
Well, that that's a hard rule.
Cannot get a refund.
Right.
Let's use the dead pet example,right?
Like if the customer expressedtheir pet passed away, well,
(15:44):
maybe we'll make an exception.
You can you're allowed to makean exception for that or
something, but you have to teachit how to make that judgment.
You can't make a human-basedempathetic judgment.
And that's a big, big one thatcustomers ask for all the time,
especially when it comes tothings like warranties, where
it's like, well, you know, youhave to make a judgment.
Was this a manufacturer'sdefect or was this a uh was this
human error?
(16:04):
I mean, I don't think humans doa great job at that, let alone
trying to teach an AI how to dosomething like that.
So that's the first one.
And then the second one is thatthere are certain things, and
this will change, and itprobably will change by the time
you release this episode, butthere are certain things that AI
is just not good at.
I saw a LinkedIn post wheresomeone asked an AI to draw a
(16:26):
map of Europe.
Uh and by AI I mean a largelanguage model, and it did a
horrible job.
It lay it was two differentFrances, uh, it labeled Berlin
instead of Germany.
Like it was just, it was awful.
It was a horrible map, right?
And he was like, checkmate AIpeople, you know, can't even
draw a map of Europe.
And I was like, but that's youknow, that phrase, I think it
was an uh Einstein phrase wherehe said, if you judge a fish by
(16:48):
its ability to climb a tree,it's gonna fail.
That's that's what this isdoing, right?
You're judging an AI based onsomething it wasn't intended to
do.
And so when we get customers orwhen I have conversations with
people about AI, that'ssomething that I always am very
specific of.
This is a large language model,and like the key is in the term
(17:08):
language, there, right?
It it's very good at guessingthe next word based on the
previous 10 words.
It's not good at certain otherthings and it will improve.
And you can see some of theproblems with the improved AI uh
image generation.
You know, there's otherproblems because of that.
But I think people need tounderstand what it cannot do
because it's it's not magic.
Priscilla (17:29):
Yeah.
Where do you see AI saving themost time for support teams?
And do you have any examples oftasks that you've seen where AI
just does a better job than ahuman and can make things more
efficient for the human to workon other things?
Craig (17:45):
So specific to support, right?
I think the AI solutions thatare in the market, you know,
have two at a high levelfunctions, right?
FAQ, question and answer stuffthat's completely repeatable.
Everybody gets the same answer.
AI is great at that and and canbasically abolish that kind of
level of question.
Yeah.
That is sometimes high ROI.
Usually that's more medium, lowROI, because those questions
(18:09):
already were quick.
You had a macro.
I think the one advantage AIhas is that it's not macro
based, right?
So it's more generative and cansound a bit more human than a
clearly pre-written solution.
The second one is theautomation of activities.
And this is where you canintegrate in your tools and
automate things that would takea human agent five, 10 minutes,
copying and pasting, using APIs,using integration.
(18:31):
The AI can call out thesethings, recognize when they need
to call out these things, andand even checking things like
formatting, you wouldn't believehow every different system uses
a different date time format,uses a different address format.
You know, if you'reinternational, you have you have
dozens of different phoneformats that you would have to
accommodate.
And then things like language,most AI solutions are
multilingual out of the box.
(18:53):
So instead of hiring one personthat sits uh you know, sits in
Italy and can speak Italian,Spanish, French, right?
You can now an AI agent thatdoes most of that for you,
translates tickets to English sothey can be reviewed by your
Englishy if that's what you needto do.
So that's kind of the firstthing that I would say to look
at when it comes to to supportspecifically.
But to benefit the team as awhole, right, that's where you
(19:16):
know you start needing torecognize what AI is good at,
right?
And AI is great at patternrecognition, understanding
anomalous data.
And this is where it's reallycool for things like voice of
customer, or very cool for likeCSAT interactions or effort
interactions, right?
Like the customer effort scorestuff, where you could take a
(19:37):
huge data set, a month's worthof tickets, a quarter's worth of
chat, you can correlate that toyou know your customer success
export tool, your is yourcustomer in red, yellow, green
status.
You can probably nowadays wherewe have all of our record uh
calls are recorded, you couldprobably inject a bunch of that
data.
You get all this data into abig bucket and say, for every
(19:57):
customer that's in a red status,tell me the top 10 things that
they're concerned about.
Yeah, you know, and and AHI isreally good at that.
Like, really, really, reallygood at that.
That's something I thinksupport can use to heighten
their visibility within thecompany, right?
By able to say, like, go to theproduct team, go to the service
team, go to the onboardingteam, go to whatever, and say,
(20:18):
Hey, I have hard evidence thatthis is the the case, right?
And so those are those arereally kind of the places I
always tell people to start.
But I mean, there's tons ofthese things, right?
I I don't reformat anythinganymore.
Like if I have if I have datain format X and I want it to
format Y, I just upload it to anAI and say, here's an example
(20:39):
of the format, the old format,here's an example of the new
format, go.
And you know, a minute later, Idon't have to worry about, you
know, changing things from CSVto JSON or or whatever it might
be, like whatever you have to dowithin your support department.
Right.
There's tons of these littlethings.
Priscilla (20:54):
Yeah.
Well, and I think you hit onone that for me feels really
huge, which is the reportingaspect of it and the help that
that can bring.
I mean, there's so much tolearn in your support inbox,
right?
And when you're working withcustomers, there's so much data
there and insight there.
(21:15):
And when you have a supportteam that's struggling to get
through tickets, they aren'table to do all the analyzing.
And so having an AI solutionthat can come in and help do
that analyzing and turn thatinto actionable results where
you can go and take that to themarketing team or the developers
(21:36):
and say, look, everyone'sconfused about the same thing.
We've got to do something tochange this.
That can be a really huge,really beneficial way to use AI,
especially if you are a littlebit more anxious about having
the AI interact directly withyour customers.
Like that's a way to kind oflike build your trust with AI
abilities, is in that kind oflike non customer facing way.
(21:59):
Also you mentionedtranslations, which is a huge
one in the way that we use AI atBuzzsprout, is that we have
people all over the world thatare reaching out and we don't
have anyone right now who ismultilingual on our team.
And so we use the translationsolutions a lot with AI.
And it's a huge help to us tobe able to communicate with
(22:21):
people on their terms in theirlanguage without, you know,
forcing them to conform to oursjust because we're speaking
English.
So I think that's anotherreally good tool, a really good
way to use AI.
What about in the regularday-to-day, you're working in
the inbox?
What little ways do you use AI?
Maybe not even directly withcustomers, but just like to make
(22:42):
your life easier.
Craig (22:43):
I mean, anywhere where you want to sound less robotic,
like instead of using a macro,AI is great at that.
So you might have a macro, butyou could run that macro.
So run the question, you know,in in the email and the macro
and say, customize this macro toanswer the question in a better
way, for example.
I think that summarizing thingsis really good.
Priscilla (23:06):
Yes.
Right.
Craig (23:07):
I feel like when you're doing a handoff of a ticket,
like, oh, I have to go onvacation, so I'm gonna hand this
ticket off to someone that'sbeen going on for a while,
maybe.
AI can summarize and say, like,here's here's what set seems to
be outstanding.
Here's the types of things thatwe're already trying, so you're
not repeating yourself.
The tone, you know, like somepeople struggle with tone.
And and I mean, that's a that'sa skill, right?
(23:28):
To to add empathy and so run itthrough and say, Hey, like, I
really need this to soundempathetic.
We made a mistake.
I don't want to place blame, Idon't want to assume fault, but
at the same time, this messagehas to be clear.
And this is the type of stuffthat's AI is really good at.
Yeah.
You know, you could write smalllittle scripts.
I I have a few uh that I usefor troubleshooting.
(23:50):
This is a great example,actually.
So we we have some logs thatstore the entire interactions
between our customers and ourAI.
And some of the details are notexposed to the user because
they're usually too technical,or there's some backing stuff
that of course you don't want toexpose to any any end user.
Right.
But we need access to itbecause if the customer can't
figure it out, we have to dosome troubleshooting from our
(24:12):
supports perspective.
The problem is that to accessit, you have to, you know, take
some piece of information,search a database to get that
information, get the export,then look through every single
line of the database to findwhere the error occurred.
And that was a time consumingprocess.
I wrote a small tool.
And by wrote, I mean I vibecoded using, you know, using
cursor, because I'm not strongat Python at all, but I wrote a
(24:34):
small script that I would giveit uh the identifier, whether
that be a keyword, what we calla conversation ID, which is a
unique identifier for everyconversation.
It goes to the database, getseverything related to that
conversation.
And then not only does it giveyou the export to a Google Sheet
of all the raw data so you canreview it.
I wrote a prompt that wouldtake all that data and spit out
(24:57):
a report on that data.
So, what did the AI find?
You know, talk about anomaliesand pattern recognition.
I would say, like, did you seeany failures in the
integrations?
You know, that's a big cause ofproblems, right?
Um, did you see any repetition?
Did you see all these things,like a list of 10 things to
identify?
And it spits out a report.
And now you could start thetroubleshooting instantly.
(25:18):
It takes about uh 30 seconds torun all that through the
script, and you have a reportthat in theory you could just
send to the customer and say,Well, here's the root cause of
the problem.
That is super powerful, right?
If you're dealing withtechnical support.
And AI is awesome at that.
Priscilla (25:33):
Yeah.
Craig (25:33):
Because it it's all pattern recognition, and that's
what AIs are good at.
Priscilla (25:37):
Yeah.
Another area that we use AI ishelp us with research.
There's a lot of support thatwe get questions about adjacent
software or platforms thatpeople are using, where
typically you would expect thecustomer support team to say,
hey, sorry, we actually don't dosupport for X over there.
(25:59):
We do support for Bud Sprout.
And so we can help you withthis, but I'm not an expert on
X.
And I'm just saying X as arandom example, just realized
that X is also a platform.
So didn't do that on purpose.
Um, but it's it's one of thosethings where you kind of feel
like, oh, I'm kind of trappedhere because I don't know how to
answer the question about thisother platform.
And that's what they're askingme about.
(26:20):
And a way that you can use AIto level up that response is to
do a little bit of researchthrough AI and say, okay, here's
the question they're asking.
Can you lead me to the righthelp article or can you explain
something to me?
And then it allows you as thesupport specialist to get a
higher level of knowledge and todeliver that to your customer,
(26:41):
which is so big when you're notworking for that company.
So you're now giving them acomplete experience with the
help of AI.
Craig (26:51):
I love that.
In fact, I actually just didthat for one of our customers.
They had a CSS problemembedding our widget on a
specific platform that theywanted to embed it on.
And I was like, I've never seenthis platform before.
I don't know anything about it.
And I did exactly that.
Yeah.
Exactly that.
It just went and said, okay,well, I can't answer the
question, but I can tell youthis is where we need to look
somewhere in this area.
And right.
Yeah.
(27:11):
Great example.
Priscilla (27:12):
Yeah.
And kind of on that same laneis brainstorming, which is
something that I really love.
I think it's probably the way Iuse AI the most in my day to
day is to brainstorm things.
I'm a very verbal processor.
So what I would love to do issit down with another person in
a room and talk throughsomething.
That's not always going to bean efficient way to solve a
(27:32):
problem.
And so being able to sit withAI and throw out ideas and get
feedback and kind of brainstormthings.
What am I missing and help fillin those gaps?
I think that's another way touse AI, especially when you're
in the support inbox and you'retroubleshooting an issue.
Say, okay, I'm here where weare, here's where we've tried.
What am I missing?
Like, what am I not seeing?
(27:54):
Those kinds of things can be areally helpful way to just use
AI in your regular, you know,daily process.
Craig (27:59):
I completely agree.
If I start so many of my AIprompts with, I'll have my
headset in and I'll be walkingsomewhere and I'll be like, I'm
in a rant for a while, and thenyou're gonna summarize what I
say into something clean and youknow, pull out what you feel
the most coherent thoughts areand most relevant.
When I'm alone in the car, I'llI'll use to write emails.
I'll just plug it in and say,like, I need to write three
(28:20):
emails.
One is about this topic, andhere's what I want to generally
say.
Next, here's, you know, and Ioften don't send them verbatim
because I find that even tone ofvoice still is a bit off, even
if you've tried to train it.
Yeah.
But at least it gives me theframework and I can just copy
paste and then change the fewwords, which I still think is is
faster.
And I can be productive, youknow, in in the car sometimes.
But you know, formattingdocuments, you know, like a root
(28:42):
cause analysis is sometimes abig one in support, depending on
your sport team, right?
You might have to deliver aroot cause analysis doc.
Don't make up a format.
Don't try to figure out thebest way to highlight things or
what's the what should go in theexecutive summary.
Upload all the data about theproblem, whatever it was that
you're root cause analyzing, andsay, build me a root cause
analysis report, and it willspit you out something that is
(29:05):
in a format.
Again, you're not relying onknowledge there.
You're relying on the fact thatit's betrayed on thousands and
millions of the root causeanalysis reports that it can
give you the formatting thatmakes sense for your use case.
And yeah, all of those usecases are ways that you can just
save minutes or hours.
And you know, going back towhat you said, people are you
(29:26):
know fearful of AI and someone'sjob is not going to disappear
because you are reusing the bestpractices on how to build a
report, right?
Priscilla (29:35):
Right.
Well, and I think that that'swhy conversations like this are
so important because the peoplewho are fearful of losing their
jobs, the people who are worriedabout the stereotype of, okay,
this AI chatbot is going toreplace me, and that is the
extent of AI and customersupport.
Hearing these conversationswhere we're breaking it down in
(29:56):
a way that's like, hey,actually, this is something
beneficial to you.
It's important to hear thesekind of conversations where it
takes it out of this like scaryunknown to, oh, okay, you're
talking about using AI as a wayto brainstorm, or you're talking
about using AI as a way to helpyou analyze your tickets.
Like those kind of things, Ithink, really help and take away
that fear factor of thischatbot is gonna do my job
(30:19):
better than I can, and now I'mno longer gonna be necessary.
Okay.
So when I think of my earliestconversations about AI here at
Buzzsprout, one of the thingsthat I was worried about was
that we had built up for thelast couple of decades this
relationship with our customersand this expectation where they
(30:40):
knew that if they emailed us,they were gonna get a response
from a human really quickly.
And so I think it's reallyimportant for us to talk about
some of the mistakes that peoplemake when they implement AI
because there is a lot of riskthere because you are working
with your customers who are, atthe end of the day, the most
important part of your business.
So, you know, what have youseen kind of as examples of what
(31:01):
not to do when people areimplementing AI in maybe not the
best way?
Craig (31:06):
Sometimes the tendency is to go back to uh decision
tree-based logic.
That's the first mistake that Itry to talk to our customers
about is the benefit of AI isyou don't have to do decision
trees anymore.
For example, if you're aclothing company and you want
to, you know, help people pickthe right size of the clothing
(31:27):
you sell, there is a set ofthings that you need to know
about the person's purchase.
If they're a man or a woman ora child.
If it's for a shirt for a man,you probably need chest size.
If it's for a uh blouse for awoman, you might need uh bust,
waist, and hips.
You have a set of things youneed to know.
And the historic way of solvingthat was to have a decision
tree.
Are you a man or a woman?
(31:48):
Man, oh okay.
Are you buying pants, shirt,you know, pants?
AI, you could just teach itthat.
Here is the things you need toknow.
Here's where that informationis.
Whatever you don't know, figureout with the customer or infer
based on the customer's profilethat you have access to or the
context that you have access to,and then only ask the questions
(32:09):
that require.
Don't repeat them, stone havethem repeat themselves.
You have context of who thatperson is, put all that together
and only ask the questions thatyou might not have an answer
to.
Using macros is a reallyinteresting one.
We have customers who have thisbeautiful generative model that
can generate reallyhuman-sounding, personalized
messages, and they want to reusethe macro text from their help
(32:30):
desk and just set use our AI tosend the same text as the the
help center would have.
And the last one, and probablyone of the biggest mistakes I
see is people, and this is athis is one that gets to me more
of a personal level, and thisis maybe my hot take for the the
podcast.
I love it.
CSAT for AI is different thanCSAT for a human.
And so many people are solocked in, dude.
(32:55):
We have to have CSAT.
We have to have CSAT.
And they forget that CSAT is isnot really what you need to
measure in AI.
Data shows, you know, if you ifyou were to look at all of our
customers and export the CSATdata, it's always slightly below
average.
And that's what you wouldexpect for a neutral, like
people use AI differently andreport on it differently than
(33:18):
they would with a human.
Because if I'm asking, do youship to Germany?
And the answer is no, that's anA and a pretty neutral
question.
And if the answer, it's eithergonna be, yeah, oh, you know,
I'm sad because, you know, it'snegative because you said no, or
it's positive because you saidyes.
Like there's no, there's nonuance to CSAT like there would
be with a human, where it'slike, oh, their tone was they
just treated me poorly, and youknow, I'm gonna give them a
(33:40):
lower score, right?
What you really want to measurewith AI is the things that
frustrate people about AI.
Did they go into a loop?
Did they ask the customer torepeat themselves?
Did the AI misunderstand in thecustomer and say, no, that's
not what I meant?
Right.
Those are things you should bemeasuring.
I call that frustration index,and that's something that our
product measures because CSAT tome is not as relevant with AI
(34:04):
as people think it is, becauseit's just a totally different
concept when it comes to howpeople aren't frustrated by
humans differently than they areby bots because they treat bots
differently than they treathumans.
Like it's it's not an apples toapples comparison.
So that's one that I alwayslooked at.
And coincidentally to yourquestion, is that everything
that I put in that list offrustrated index, I would label
(34:26):
as the mistakes you do when youimplement AI, right?
If you allow your AI to getinto loops, if you don't train
it on synonyms well enough sothat it understands the context
and and and misunderstands yourusers, you know, if you don't
have the right guardrails inyour product that prevent
hallucinations and sell you uh,you know, a one-dollar car, like
that Chevy example or the Idon't know if the Air Canada
(34:46):
example where they gave away abereavement flight that didn't
exist, you know, made it.
But that's that's the problemwith AI.
That's bad AI.
Yeah, right.
It's that those are themistakes you make, is that you
don't put those guardrails inplace.
So those are kind of the maybethe top three or four examples
that I have of people like, youknow, implementing AI poorly.
Yeah.
And that you need to avoid.
Priscilla (35:05):
And I'll add my own to that, which is not giving
people a way to get to a human.
Craig (35:11):
100%.
Priscilla (35:12):
I think that a lot of times I will hear these stories
of people like jumping into AI.
It's so exciting, it's sogreat, but they don't give you a
route out to a human.
And I think that's a hugemisstep to not have that,
especially if your users areused to working with humans.
And then now you put this AI infront of that, you're gonna
lose trust if they don't have away to get to you.
(35:34):
And I think you said itearlier, this idea that, you
know, it's not that they don'twant to work with AI, they just
don't want to work with bad AI.
They don't want to work with abad chatbot that's gonna throw
them in the loop.
And so you're gonna run intothose situations.
And I think that frustrationindex is a really good way to
measure the success of your AIis okay, how much are my
customers getting stuck in thesefrustration cycles?
(35:57):
Because that's what you want toavoid when you're implementing
AI.
You wanna make sure you'regiving a smooth, top-notch
experience because I thinkcustomers are gonna have a
shorter fuse when it comes toworking with AI than with a
human when you're like, givethem a bad response.
Immediately it's gonna be like,nope, I knew I knew this was
gonna be bad and now I want out.
You know, so you have to bereally careful with that.
(36:20):
Do you have any funny storieswhere AI just got it wrong?
Craig (36:23):
I would say it's wrong, but I have a couple of funny
ones where we had a particularclothing brand and it was very
flowery, very summary, like thatwas their brand.
And they speak with the poetic,you know, nature of Maya
Angelou was something again,very kind of like poetic and
summary and happy.
And so when we were buildingthe tone of voice, um, we put
that in there, you know, speakwith the poetic, you know,
(36:46):
angles of of Maya Angelou.
The bot was hilarious.
There's some of my favoriteones that came out of this was,
you know, it was things like,you know, what's your shipping
policy?
It's like to traverse the worldwith through the winds of
change, as the, you know, and itwas just it went on and on and
on.
It was like, we will take it toEurope, we will take it to
Asia, and we will even cross thePacific over into the Americas.
(37:10):
And like it was just that wastheir shipping policy as
interpreted by that.
And then we had anothercustomer, they were uh a company
owned by a famous uh rapper,and so they wanted to have that
kind of built into the samething in the tone of voice, and
and the same thing.
The the the questions came backwith this like real beat, like
you know, kind of nonsensical,lots of slang, lots of so those
(37:32):
are a couple of funny things,but I can't think of an example
where an AI has gotten itcompletely wrong.
I think there's some reallyinteresting things that you
don't think about when you'retraining AI.
So, for example, we this wasactually a Kodif example.
So we use Kodif on our websiteas you might expect, and someone
was asking for a phone numberto contact us, and we don't
really have a receptionist, likewe don't have a phone in our
(37:53):
office.
We're a small startup, but theAI, in one of our privacy policy
documents, found a phone numberfor an old office when the
company was first incorporated.
So the document was out ofdate, but but it found this
phone number, and it's like, oh,the only number I have access
to is this phone number, whichthe person then called, and of
course it was disconnected.
And so it was one of thosethings where it's like, well,
(38:14):
where did it find the number?
But it was just so funnybecause someone was like, Well,
how did it, where would it havepicked up a number?
And it did it hallucinate thisphone number, but it didn't
hallucinate it.
It was it was in the trainingthat, oh, it had said if you
looked at the document tocontact Kodif for uh a privacy
or security concern.
Here's this phone number.
Yeah, just stuff like that,where we get these customers who
(38:36):
who ask us, well, how did itcome up with this response?
And you look and it's like,yeah, you know what?
It actually it makes sense itcame up with this response
because of these uh theconnections it's making here,
even though those connectionswouldn't wouldn't have been made
by a human anyway.
I love the I love the TotalVoice things because I think
that those are reallyinteresting interpretations.
Priscilla (38:56):
Well, and I think it speaks to the importance of, and
you were talking about this alittle bit ago, but the idea
that like our skills are goingto change now for customer
support specialists, becauseyou're gonna need to lean into
these other skills in order tocorrectly train the AI and those
strategies.
And so it kind of leads wellinto my next question, which is
(39:18):
about those skills, because youneed to be able to have people
on your team who can train theAI in the right tone to navigate
those kinds of um examples thatyou just gave, where someone
asked for a certain kind of toneand it didn't react in the way
that is gonna ultimately be theway you go moving forward.
So, how do you train people ontraining the AI?
(39:38):
So, what other kind of skillsdo you see needing to be
developed?
Say you're in customer supportand you're looking at the future
of your career and you're like,I want to invest in some of
those skills now.
What are those skills thatpeople are gonna need to have
sharpened going forward incustomer support?
Craig (39:56):
So, I mean, that the number one is just understanding
how AI prompts work.
Priscilla (40:01):
Yeah.
Craig (40:01):
Right?
AI prompting is a is going tobecome a skill.
And in fact, I would argue it'ssomething that they should
start teaching in in the highschool areas.
Yeah.
It's gonna become thatfundamental.
And again, in five years,probably I'll look back on that
statement and say how stupid Iwas, because now AI can
interpret anything.
But yeah, right now, AIprompting is so vital.
And what I mean by that is thatlet's go back to the root cause
(40:22):
analysis report example.
In a world where you'reproducing that report, you know,
the skill is to logically takefactual data and analysis and
put it into some logical orderto read, right?
That's the skill.
In a world where you're notproducing the report, you are
sharing the information withinwith an LLM and asking it to
produce the report based on theinformation.
(40:43):
The skill is to describe theinformation, what are the key
factors it has to highlight,what are the things to ignore,
what to point out, what not topoint out.
And there's a specificity tothat, right?
There's a skill to being ableto be very precise in your
language to tell something howto do something.
And uh the better you're at atdoing those things, the more the
(41:06):
report will sound like you, themore the report will look like
how you want it, the lessediting you'll have to do after
the fact.
But that bit is going to becomethe skill.
And I think the number twoskill is we have to get into a
mindset of this personalizationmore as humans, too.
I had a really cool experiencewith Air Canada recently where I
had to change something on aflight and I called, and the
(41:28):
first thing it said to me beforeit even started ringing was
based on your phone number, uh,we think you're calling about uh
your flight to Denver nextweek.
If this is correct, press one.
Wow.
Right?
Like, wow, seriously, becauselike now, how many times have
you tried to say those six-digitalpha numeric codes in a broken
NATO alphabet?
And N, you mean N as in Nancy?
(41:50):
No, M, M, N, you know, yeah.
That was a beautiful use of notreally even AI, it's just use
of technology to get that andpersonalization.
Personalization, but becausethat's what people are going to
get used to with AI, humans haveto adopt that mentality now and
be like, oh, if I have thatinformation available, if I if I
see a phone number with a name,or if I see uh the person has
(42:13):
you know purchased this oneproduct recently, again, don't
ask them to repeat themselves.
Don't ask them to spelleverything out for you.
Use the context to help them.
Start the conversation with uh,well, how did you like that
lipstick that you bought lastweek?
You know, like there's a way toadapt that into your style and
tone of voice.
And I think again, that's askill you're gonna have to teach
(42:33):
uh your team.
And I don't think it'sunreasonable to say that that
will become an expectationbecause if a bot can do it, a
human should be able to do it.
Priscilla (42:41):
Yeah.
Well, I think it's one of thecoolest things about AI and
customer support, is it's goingto give customer support
specialists the ability to geteven more personal and to
personalize things and to focuson building relationships
instead of just closing tickets.
And I think that's really whereif you're incorporating AI into
your workflow so that you canfocus on creating stronger
(43:05):
relationships with yourcustomers, you're doing that in
the right way.
Right.
You're you're looking at itfrom the right perspective.
If you're incorporating AI toeliminate a support team, that's
probably not the right route totake.
In five, 10 years from now,when we're using AI still, what
aspects of this customer supportrelationship environment, like
(43:28):
what are the things that need tostay human?
Craig (43:30):
Things like empathy and situational re uh judgments and
reactions are always gonna haveto be in a human form, right?
Yeah.
It's so hard when you say five,10 years, uh, because I just
this this technology is movingso fast.
So fast.
It's one of these things that Istruggle to know what my seven
year old is gonna do in theworkspace, right?
(43:51):
Not because I don't think he'sgonna have a job, by the way,
just as we already talked about.
But so when it comes to whatshould stay home, then I think
it's a stuff that is uniquelyhuman.
Situational reasoning, I think,absolutely is there.
There's there's no, you know,this artificial general
intelligence, the AGI that thepeople talk about.
I'm my God is it's not five, 10years away.
Like we're it it just doesn'tseem possible to me.
(44:13):
I don't think it's a hundredyears away, but I, you know,
creativity, like solving aproblem.
You you mentioned, you know,you do research with with AI,
which again, I do too.
I think it's a smart idea, butthat's not ever going to
generate a new idea.
It's gonna validate otherideas, it's gonna say, hey,
other people have done it thisway.
It's gonna help you absolutely.
Yeah, no criticism, but it'snot gonna generate a net new
(44:34):
idea.
It can't, because that's nothow AI currently works.
Right.
AGI would have to exist forthat.
And so that's the stuff.
Anything, you know, being ableto, again, from a support
perspective, build a creativesolution, like, oh, I'm sorry,
we're not gonna be able toreship this thing because it's
out of stock, but here's what Ican do because you know, I
you're a valued customer, right?
(44:55):
Like that type of creativity,uh, of problem solving.
And I think anything to do withreally deep product or service
or whatever industry knowledge,like that deep level where it's
like, I understand exactly whatyou're trying to achieve with
our products and services, andI, as an expert in that area,
(45:16):
can help guide you down a pathto success for this specific use
case.
Again, that that's justsomething that if you, you know,
you might be able to train anAI on on some of that, but I
don't think it's gonna have thatdeep, full understanding that a
human conversation can have.
And so, yeah, those are areasthat I think are uniquely human
that that probably won't go awayuntil AGI exists.
(45:38):
And even then, AGI is not magiceither.
Yeah.
It still has to have this basisof how to operate.
And I do picture a world whereI call it the big white box.
That's my my name for it.
Everyone is gonna have this bigwhite box, whether it be a
website or or maybe it's aphysical box.
It doesn't really matter.
And that thing, that box knowseverything about you.
(45:59):
It knows your age, your height,your clothes size, it knows
what type of dishwasher youhave, it knows where you live,
it knows that your floors arecarpet, it knows whatever.
It knows everything.
And when you have an issue,you're gonna go to that big
white box and describe yourissue, and it's just gonna
figure out how to do it, whetherthat be contact a human or a
sport department, whether it beto call a service person to come
(46:22):
and fix it, whether it be tobuy something net new, whatever.
The box will know everythingabout you, will know your
spending habits, budgets, itwill know everything.
And you just say, you know,hey, my upstairs tap is leaking,
and it's gonna be like, oh,well, that tap is still under
warranty from Mowen.
I'll submit the warrantyrequest, and here's the place to
do it, and I know your address.
(46:43):
So I'm you don't even have tofill it in.
It'll be there tomorrow.
That's coming, right?
Yeah, so that's a differenttype of support, right?
That is someone, you know, fromevery all these companies
plugging into the big white boxto say, you know, if I want to
be a plumber and I want aservice call to these houses,
that big white box needs to knowabout my plumbing services and
my service radius and my pricesand my and so there's another
skill.
(47:03):
You know what I mean?
Like it you're gonna have toteach this big white, and that's
but that's where we're going,right?
And and so the you know, goingback to what the human element
is of that, it's the teaching ofthe big white box, how to get a
hold, how to start a warrantyrequest for a dishwasher, how to
hire a carpet cleaning servicebecause you spilled wine on it.
Like that's the thing.
That's the next you know,another layer of how support's
(47:25):
gonna get there.
Priscilla (47:25):
Yeah.
I feel like we've touched on somany things in the last you
know, hour.
I'm gonna ask you one lastquestion to wrap this up.
What is one truth about AI thatyou want people to take away
from this?
Craig (47:37):
It would be disingenuous of me, especially someone who's
in the industry, that it's notscary, right?
There's a lot of unknowns,there's a lot of things that
could go wrong with it.
You know, there were peopleback in the 90s.
I remember people saying thisinternet's gonna destroy
civilizations, and and to somedegree, maybe it has.
Maybe it has.
We can we can have thatargument as well.
(47:59):
But the thing is, is that it'sit's all about adaptation.
We should be trepidatious.
I just don't think we should beuh uh A, completely ignoring
it, and B worried that it'sgonna cause mass global
unemployment.
Uh that's just not so the truthto me is that there's a middle
ground where we accept that thisis a technology that now
(48:20):
exists, it will never not exist.
And so, how do you want to useit in your daily life?
And everyone can choose how toenable it or how not to enable
it.
And that's that's really got tobe up to up to you.
And then the rest is up to ourleaders to to help society
thrive with it.
So, yeah, that's probably my myopinion on the truth of AI.
Priscilla (48:39):
Well said.
Thank you so much for beinghere, for sharing with us,
Craig.
This was really informative.
I hope that our listeners areable to take some of this into
their daily work.
If they want to learn moreabout you, and if they want to
learn more about Kodif, how canthey learn more about you and
Kodif and how they could use it?
Craig (48:57):
Well, I'm LinkedIn.
That's the main place I am.
So look, I'm the only CraigStoss on LinkedIn.
I'm almost guaranteeing you ofthat.
Nice.
Um and uh and then Kodif is isk-odif.io.
But yeah, I mean, please reachout.
I love having theseconversations.
So if you have questions, if ifyou want to discuss something I
said or you know, if youdisagree, I I love hearing
(49:19):
disagreements too, becausethat's how we we build better,
uh, better content for people.
So yeah, that's uh that's howyou get a hold of me.
Priscilla (49:26):
Yeah, and it's good to have these conversations
about AI and to learn more aboutit.
And that's the way thingsbecome less scary, is when you
understand them more.
So if you are listening to thisand you're scared about AI,
reach out to Craig, have theconversation with him, and
you'll learn more and it'll takesome of that fear away a little
bit.
If you have a question or atopic that you would like us to
cover on a future episode, youcan shoot us a text using the
(49:49):
Send Us a Text link in thisepisode description.
As always, if you like theepisode, please share it with
someone who works in customersupport or leave us a review on
Apple Podcasts.
Thank you again to Craig Stossfor joining us on today's
episode.
Thank you for listening.
Now go and make someone's day.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Bobby Bones Show
Listen to 'The Bobby Bones Show' by downloading the daily full replay.