July 8, 2025 • 43 mins
How should customer support teams use AI without losing the human touch?
Priscilla Brooke sits down with customer success expert Conor Pendergrast to explore the evolving role of AI in support workflows. From using AI as a smart assistant to drive efficiency, to building better customer relationships through intentional automation, Connor offers real-world examples, clear do’s and don’ts, and practical tips for teams of all sizes.
Learn how to strike the balance between AI tools and authentic human service, and hear why empathy, intentionality, and trust remain at the heart of excellent support even in an AI-powered world.
Learn more about Conor at CustomerSuccess.cx and subscribe to Conor's daily tips for CS!
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!
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 customersupport from the people at
Buzzsprout.
I'm your host, Priscilla Brooke.
Today we're tackling AI andcustomer support.
Ai is all the rage, but how canwe use it in support to make
our work better, provideexcellent experiences and build
strong customer relationships?
Thanks for joining us.
Let's get into it.
(00:21):
We launched this podcast a yearago and I knew when we launched
it we were going to have to talkabout AI.
Yeah, it was going to happen.
It has been such a hot topicrecently and I was always kind
of like oh, I don't know, whatdo I want to say about AI.
I'm still figuring out what Ithink about AI and how to best
use it.
Yeah, on today, who has somereally great insight into how
(00:43):
you can use AI with yourcustomer support teams to really
continue offeringcustomer-focused support and
using AI as a tool to help youdo that.
So joining us today is ConorPendergrast.
Conor is a customer support andretention specialist with over
15 years of experience.
He has worked as a director ofinformation and communication,
(01:05):
spent 10 years with Expens as adirector of information and
communication, spent 10 yearswith Expensify focusing on
customer success and retention,and currently he is the founder
of customersuccesscx, where hehelps CX leaders grow their
careers, empower their teams anddrive meaningful business
impact.
He also writes a wonderfullyhelpful daily newsletter for
support professionals.
Thanks for joining us, Conor.
Conor (01:24):
Thanks, Priscilla.
That was quite the intro.
I'm impressed with myself.
Priscilla (01:29):
You should be.
I'm impressed with you, andeveryone listening should be
impressed.
We always like to start ourepisodes with a lot of
positivity, and so one thing welike to ask our guests is who
has made your day recently?
Conor (01:41):
I'm going a bit left field here because it has
absolutely nothing to do withwork or anything like that I
will shout out two of my son'sfriends' parents.
So I live in Birmingham theBirmingham that's in the United
Kingdom, not the Birminghamthat's in Alabama for anyone
who's in the United States ofAmerica.
Priscilla (01:56):
Thank you for the clarification.
Conor (01:58):
This is going to surprise you.
There are places outside of theUnited States that have the
same name.
There actually was a York, justa York existed, so anyway.
So I live in Birmingham and wehave the Botanical Gardens here
in Birmingham and it's a lovelyplace.
It's got like nice glass housesand flowers and it actually has
a really nice playground, andso my son's friends' parents
(02:19):
were going there and they justthrew it into the little
WhatsApp group and said, hey,we're going going to go here,
and it turned out to be a reallylovely day.
We rocked up at about 10o'clock, which is when it opened
.
It was sunny.
When I say it was sunny, I meanit was like this is england and
it's april and so you don'treally get consistent dry, warm,
sunny spells, but we did forthe whole day and I mentioned
(02:40):
earlier in the green room that Ihave a little bit of sunburn,
so that's how I got sunburned,but it was just it was just, it
was me, it was my kids, it wasmy son's friends and their
parents, and it was just like areally chill day.
And then, after we finished, weall came home.
I went for a short run, cameback I'd abandoned my son again
at their house.
I came back to collect him andthey were sitting out the front
(03:03):
drinking wine.
So then we segued straight intojust sitting out the front
having a nice drink, gettingmore sunburnt, getting some
pizza and then falling into bed,all of us totally exhausted on
the Sunday.
So it has nothing to do withwork, but I will shout out to
Eleni and Christina, who are thetwo parents that's so great.
Priscilla (03:20):
It is such a good reminder to us to you know,
think of other people.
Reach out to them Like younever know what someone else is
doing with their day.
They might be totally free andable to go and have a fun Sunday
afternoon.
So, yeah, I love that.
Conor (03:32):
Exactly.
Jordan (03:33):
And I'll tell you what when you're a parent and you
have another parent supportsystem where they're like yeah,
the kids can play, you can justdrop them off, go do your
errands.
That is so rare and preciousand you just hold on to those
people as tight as you can.
It's such an amazing thing.
Conor (03:54):
Yes, I agree.
Priscilla (03:56):
Thank you for sharing that, Conor.
Before we get into the world ofAI and how we can use that, I
know I gave a lot of yourbackground in the intro AI and
how we can use that.
I know I gave a lot of yourbackground in the intro but for
anyone who isn't familiar withyou, what else is there for
people to learn about?
What it is that you do and thework you do in customer support
and how you've kind of gonethrough that the last 15 years?
Conor (04:16):
I guess my earliest customer support interactions
where I was doing like paid work.
I worked in a shop that soldsheet music.
So when you're a musician, youhave sheets of music and you
read off them.
And let me tell you, dearlisteners, I was neither a
musician nor really like asalesperson, nor, let me just
say, I was not a good fit forthis role.
But thank you, liz.
(04:36):
Liz was a friend of my mum'sand Liz just gave me the job and
I was awful at the job.
I didn't know anything aboutmusic.
People would come in with themost basic questions and I'd
just be like I don't know.
I got to go and ask Liz, but itwas great.
But I would say, looking backat myself at that point, that
was very much like cutting myteeth on being in an environment
that I didn't know very welland didn't do great at it.
(04:59):
But I'm very proud of the workthat I do now.
It was a far stretch from whereI was.
So after that that wasuniversity.
That was the usual thing ofjust like make a couple of quid
on a Friday, spend it all onFriday night, make a couple more
quid on the Saturday and keepsome of it after Saturday night.
That was the goal.
The goal was like always end upin slight profit.
After you hit the pub on Fridayand Saturday night, it was
(05:27):
Dublin.
All right, I am Irish, so whatcan I say?
After that, I worked for asocial care company.
We created a software productin there and I spun it out as a
startup, moved to London, livedthe London startup dream, which
is mostly like getting trains tothe rest of England and having
terrible Wi-Fi on those trainsand really considering what I'm
doing in my life, did that for ayear and left and joined
Expensify.
I initially joined Expensify,funnily enough, as a sales
(05:50):
generalist, because what Irealized from that earlier role,
that startup role, was that myskills as a salesperson were
lacking, so I needed to work outhow to sell, although I have
different philosophy aroundsales now compared to what I was
expecting to want there.
Within a week of joiningExpensify, we all realized that
there was this new team that wewere creating, which was much
(06:11):
more about account management ina long-term relationship
building way, and so I joinedthat team as one of the first
four people on the team, andthen we evolved that over time,
I eventually, after nine and ahalf years at Expensify was the
senior leader there, which isthe sort of like step below
director Left there in February2024.
And since then I've been workingindependently.
(06:32):
I'm a coaching consultant forcustomer support professionals
and leaders.
So if you have, if you providecustomer support, my job is to
help you provide better customersupport.
And I do that, as I said, boththrough coaching and consulting
customer support, and I do that,as I said, both through
coaching and consulting.
Priscilla (06:45):
So since this is an episode about AI, I figured I
would let AI ask a question tokick us off.
Conor (06:50):
Oh God.
Priscilla (06:58):
Love it, yes, so I asked AI what AI would like to
know about you, and this was thequestion they asked what's one
experience you've had, either asa customer or working in
support, that made you thinkthis is why I love what I do?
Conor (07:07):
So I don't think there's a single experience that really
showed me or reminded me likethis is why I love what I do.
But there was a pattern oftypes of customer support
interactions that I really liked.
Even after nine, 10 years atExpensify, even as a senior
leader, I was still managingsome escalated support
interactions.
And it was the pattern of likeone customer hit a really
(07:28):
peculiar bug and like it wouldtalk on my little curiosity and
I'd be like I there's somethingthere, there's some pattern
there where I think there's somethat I can work out what's
going on.
And you'd like dive into thelogging tool and you'd find that
one weird error and you traceit back to the code base and
you'd be like hang on, this filehasn't been changed in six
years.
How is it possible that noother customers have ever hit
(07:51):
this error?
But for some reason it's likeit's just a bug that has existed
for six years and no one hasever hit it.
We've all seen weird upgradeflows or migrations, or you've
added a new piece offunctionality but you've needed
to change stuff in the back endto support that and because of
that sort of ingrained knowledge, you have just had this like
(08:12):
habit and pattern of how yousleuth on things.
So it's so sort of like weird.
One person will hit this in inin a six-year period.
Bugs that I, I love and I findjust endlessly fascinating and
it'll usually be one of thosethings where, like, I'm the only
person who's patient enough tospend an hour on it.
And that's all it usually like.
It doesn't take a huge amountof work most of the time, but
(08:35):
it's just.
It's kind of fun.
I like it.
Those are the fun ones.
Priscilla (08:37):
Yeah, the investigation.
I love that too.
I love the okay.
Something's going on here.
I know there's an answer.
An answer exists.
Conor (08:46):
How do I find it?
Yes, an answer exists to mostquestions.
Priscilla (08:49):
So let's get into it.
Ai and customer support it'sthe hottest topic right now.
Everyone is talking about it incustomer support, outside of
customer support, just if you'reworking in a professional
setting.
Ai has come up, and so I feellike people can hear AI and
start to worry about theslippery slope of what that
means if we make everythingfully automated and fully AI,
(09:11):
and so I want to start byputting it into a really
practical frame.
So, Conor, how do you use AI inyour day-to-day that makes your
life better.
Conor (09:20):
So, as I mentioned, I coach customer support leaders
to help them provide bettercustomer support.
That will typically involvelike an hour long conversation
every two weeks and one of theways that I use AI in that
process is I have a callrecording tool and the call
recording tool records the calland then gives me a transcript
afterwards but also gives me asummary afterwards.
(09:41):
It's that kind of like reallysimple follow-up step that then
I've dropped that into theNotion page that I've created
with the client and so we canboth go back.
So the following time when it'stime for us to speak again the
day before we speak, I willalways go back and I'll look at
the action points that we pulledout of that.
I'll look at the summary.
I'll look at the video if Ineed to, but very rarely.
(10:10):
The other thing I started doingthis week as well, hot off the
presses this week, was I pulledthe transcript and I ran it
through chat, gpt and I said toit, like what priorities should
I have going into the nextconversation with this person?
And it pulled out things and itpointed out a couple of things
like oh, here's what they saidthey wanted to do and here's an
approach that they said theywanted to avoid.
Here's the style that theywanted to take.
And because I havepre-programmed ChatGPT with my
own set of instructions, itdoesn't do like wishy-washy
(10:34):
stuff.
I've told it to be very directwith me.
You probably know the Kim Scottbook, radical Candor I told
ChatGPT in the custominstructions I said use the
Radical Candor approach.
So every time now it sayssomething to me, it says I'm
going to challenge you on thisspecific point.
And it doesn't say like here'sa gentle nudge or anything like
that.
I hate the gentle nudges, I donot like gentle nudges, I just
(10:55):
want clear and direct feedback.
And so it gave me the prepnotes.
It said like here's the thingsthat you might want to consider
bringing up again.
And then I went into theconversation and then I fed the
transcript from the secondcoaching session into it.
I said, ok, did I hiteverything?
I'm pretty sure I hiteverything.
And it pointed out some placeswhere I had failed in my role as
(11:16):
a coach.
It pointed out places where I'ddone really well and like
clearly pushed this person andclearly like directed them to
make progress and make decisionsthat they were not making.
That's one small way in whichI'm using an LLM like an AI tool
.
The other ways I use it.
I do use it in podcast prep alittle bit as well.
So if I want a quick summary of, like, for example, topics that
(11:37):
haven't come up, gaps thatmight be interesting to talk
about other podcasts that aresimilar to podcasts I've been on
before all those kinds of waysthat you can just save a little
bit of time, a little bit ofGoogling.
That would speed it up.
I use it to rewrite things,sometimes to get it closer to my
voice a little bit, or to like,change it from the style that
(12:00):
I've written an email in to aLinkedIn post, for example.
What I will do personally isI'll avoid using AI to create
things itself, but I will leanon it, like I said, to challenge
my ideas and to give me directfeedback instead.
I think that's a useful thingfor me so far.
The funny thing is it's likethis is all very much in the
context of it's useful forsomeone who's a manager,
probably, or a leader.
(12:21):
But, if you're on the frontlines, if you're working
directly with customers, thisisn't how you're going to be
using it, but there are probablylessons you can pick out from
it as well.
Priscilla (12:30):
Yeah for sure.
So, as someone who has been inthe industry for 15 years, how
have you seen AI really changethe way people approach customer
service, especially in the lastfour to five years, or really
like two to three years, as it'sbeen really hitting the
mainstream and becoming used byso many people?
Conor (12:48):
Yeah, I would say even two years ago, I was extremely
skeptical of how effective AIproducts could be in managing
customer interactions, mostlybecause it seemed really basic
and, a lot of the times, like aperson could do it better with
better information and bettercontext, faster really.
And it's really only in the lasttwo years that I've started to
(13:10):
change my mind even morerecently than that, I'd say just
in the last year of seeingactual progress on AI based
tooling for customer supportprofessionals.
And there are real downsides tothat.
Like I think there have beensome really destructive actions
that some companies have takenin the last year fueled by this
idea that ai can do all of theircustomer support flows and it's
utter, utter rubbish and uttergarbage.
(13:31):
Like I think the real valuethat we're starting to see is
that you can use ai as a tool tobring great customer support,
but that will be by allowingyour support team, and your
success team as well, to engagein higher value, higher leverage
activities or just higher valueactivities in general.
We as humans, are creative, weare insightful, we are
(13:54):
interested, we are curious.
We are the ones who areactually creating these
companies and creating theseproducts and services, and we
are going to know them the best,and we should use AI tools to
support us, not vice versa.
It's just a baffling idea to doit the other way around.
Jordan (14:11):
Yeah, I actually saw a story recently where an AI
customer service bot had fired acustomer and they were just
like it's best for you to cancelyour subscription because I
don't see a route that you cantake.
I was just like, wow, it's bestfor you to cancel your
subscription because I don't seea like a route that you can
take.
I was just like, wow, that'srough.
Priscilla (14:30):
It's dangerous.
Yeah, it's interesting becauseif you look at like support
forums or I know we're both insome different support
communities and on LinkedInyou'll see there's this push and
pull of people who are reallyexcited about the tool AI and
how they can use it to maketheir work better.
But then you'll see people whoare nervous and they're anxious
(14:50):
about what that means for thejob they have and the value they
bring to a company, becausethey've seen things like what
you were talking about and justthese massive layoffs because
people feel like the AI robot orchatbots can fully do what
their role is and the value theybring.
So what would you say to thepeople who are in customer
support right now, who are kindof in that worried stage so
(15:12):
maybe they're not quite ascomfortable with AI yet they're
still like I don't know, I don'tlike this, I want to fight
against it what would you say tothem?
Conor (15:19):
Yeah, I think we have a real challenge in that we do
have sort of like a perfectstorm of really like not great
situations happening in a lot ofways, and customer support has
traditionally not been very wellvalued by companies and it's
been really under-resourced andthese are already unstable
positions in a lot of cases.
I think, broadly speaking, yourwork will look dramatically
(15:43):
different, but I think you willstill have a value to bring to a
company, so long as the companyvalues you already.
Priscilla (15:50):
Right, and don't just completely ignore that AI is
there and hope that it goes away, because I don't think that's
happening.
So what customer support tasksdo you think AI is best suited
for?
What would you give to AI andwhat would you say, oh, I don't
want AI to touch that.
Conor (16:05):
Yeah, it's a tricky question.
The first one is like if thereis a reliable association
between a question and an answer, you can probably have AI
engage in that.
That could be like you have asolid knowledge base great, good
work, you set yourself up forsuccess.
You should have had that anyway.
But now there's like a bigincentive to actually have a
great knowledge base.
(16:26):
It's like before now it waslike oh, let's build a help
center.
Yeah, Okay, great, we threw onetogether, It'll work fine.
And now if you have a wellstructured health center, you're
kind of off to the races andyou can do really great work and
get your AI doing that becauseit's got a solid basis of
information and when itcontradicts what you think, it's
going to be pointing out thatyour knowledge base is wrong.
(16:48):
If you don't have a solid baseof any knowledge, then you're
not really going to get muchbenefit from AI, because it's
just going to be making stuff upand you don't want that.
If you can't afford a wronganswer, you probably shouldn't
get AI involved either.
So that might be, for example,for topics around compliance or
(17:13):
topics around legal questions oranything like that, where you
don't want it sort of firing offsome half-thought-out answer or
its best guess of what it is.
You could probably use AI insummarizing and interpreting and
suggesting, but a human expertshould probably be involved in
reviewing and confirming thosesort of things.
And then there's the likesensitive topics as well.
If you have a product that hasa subscription associated with
(17:34):
it, cancellations and discussingcancellations you probably want
that to go to humans a lot ofthe time as well.
A lot of people will be verycomfortable with having AI
interactions.
Just do the whole cancellationflow for them.
If you want to go down theroute of fully understanding the
problems and seeing if you canrescue an interaction, then you
(17:54):
probably want to have that gothrough humans as well.
Priscilla (17:57):
Yeah, and I think the holiday after hours coverage.
I think that's another way thatAI can really be helpful,
especially if you're a smallteam and you're struggling to
provide coverage for what yourcustomers need or struggling to
take time off, letting AI comein and stand in the gap for you
so that you can have a break andletting AI take that.
(18:19):
And, like you were saying just asecond ago, I liked the idea of
not giving AI the work thatcannot be messed up, kind of the
idea that you know, if it'sthat important, it really should
be someone who can sit thereand give it the focus that it
needs, because with AI, youmight not have a quick review of
what happened and so somethingcould slip under, and so I think
(18:42):
both of those are good examplesof when to use AI, and then
some ways where you know reallyyou might want to be human first
in that situation.
Conor (18:51):
Exactly.
There's also opportunities forlike decisions that are really
about your company and yourcompany's culture.
So one topic that came up thisweek with the client was
thinking about like how, whenyou offer premium support, how
are you framing your premiumsupport?
Should it always go to humansupport agents or should it also
go through your AI agent first?
(19:13):
In that particular case, I wasof the opinion that the average
outcome would be better if itgoes through your AI interaction
first and then, if it's notaddressed, then it just gets
quickly escalated and it's aprioritized response at that
point.
But that's going to be veryparticular to the type of
support that you're giving andwhat the expectations are as
well, because if you're doingpaid support some people might
(19:38):
expect that you always getthrough to a person.
But like that means you have tostaff like 24-7, 365.
That's going to be an expensivepremium support option and if
someone's paying a relativelysmall amount per month for that
premium support option, it's notgoing to be human.
First, and it probablyshouldn't be, because then it
would be a lot slower and peoplewould be a lot more frustrated
by that as well.
Priscilla (19:58):
Yeah, that's a good point.
We don't do premium supporthere at Buzz Brown.
It's something that I don'thave a ton of experience with
myself, kind of having tiers tosupport.
But I can imagine that if I waspaying for this premium tier,
that I wouldn't want an AI to befirst.
But at the same time, if it's anight or a weekend, there's got
(20:19):
to be some understanding that,hey, there might not be a person
ready at my beck and call on anevening or a weekend.
Now, maybe during working hoursI always have a quick response
from a human.
But you might find that there'ssome times when that AI first
can still be the initialexpectation.
So one of the things that we doon this, we're saying, hey, we
(20:54):
really care about human customersupport and feel very strongly
that that is what makes youstand out in your competition is
when you have these humaninteractions through customer
support.
So where does AI live in thatworld of really human,
empathetic, personal customersupport?
Conor (21:14):
God.
It's such a curious topicbecause relationships are
actually created in a whole hostof complex ways and everyone
listening today will haverelationships or feel like they
have relationships withcompanies and they don't know
the people who work there.
And it's the strangest thing.
(21:34):
What I'm talking about iseffectively brand loyalty or
maybe, dear listener, parasocialrelationships, like parasocial
relationships, have become thissort of buzzword.
Talking about podcasts and Ilisten to podcasts I listen to
podcasts all day long.
We do too.
It was the London Marathonrecently and I was following a
(21:54):
couple of people I actually know, but a couple of more like
better known athletes.
It was funny.
Afterwards, I was equallyinvested in knowing how Claire
got on, as knowing how Phillygot on, even though Philly is a
YouTuber and like a runningprofessional athlete like she's
a proper professional athletebut like see videos from her
(22:15):
once to twice a week and Ireally want to follow up on her.
It's not from her having arelationship with me, it's very
one sided.
It's entirely based on what Iam feeling and how I am engaging
with her as a creator, and so Ithink, actually, thought
(22:45):
through that lens, you canactually use AI to promote
relationships between yourcustomers and your company and
products and services as well,because it's effectively just
another way of having consistentbrand perceptions and
consistent brand interactions.
Brand perceptions andconsistent brand interactions,
so long as you have set it up ina way that it all acts
naturally together and feelslike it's all part of a single
ecosystem.
I think that you will be able touse AI as a way of building
relationships with customers aswell, which is sort of the
(23:06):
opposite of what I would everhave said before.
I started thinking about thisand I will say AI doesn't have
empathy.
It's not real.
A lot of the writing that AIdoes it winds me up so much.
The sugar sweet kind of overapologetic stuff.
That's not what I want.
As I went back to earlier.
I like clear, direct and kind,but to the point, without being
(23:29):
over sweet.
So yeah, that's my thought, butlike I'd love to get your take
as well.
Priscilla (23:34):
It's hard, right, it's an interesting question,
right, it's the idea that I wantto use this artificial
intelligence to help buildrelationships with my customers.
But this is not a human person.
But I care about the humaninteraction of myself, as a
person who works for andrepresents Buzzsprout, with the
customers who use Buzzsprout.
(23:54):
I think, personally, that thatis a huge part of offering
really remarkable customerservice, and so when you bring
AI into that, you go okay, well,those things can't coexist,
because one is a robot and oneis a human and you can't have
human interactions with a robot.
It just can't happen.
But I think the thing that youknow you were talking about with
(24:17):
the person you follow andwanting to know how they did in
the marathon Well, the reasonyou care about them is because
they have been authentic inshowing you their lives and
bringing you into their lives,so you are the one with the
feeling and the connection tothem.
They don't have a connection toyou, but they have opened up
themselves so that you canconnect with them.
Conor (24:39):
Yeah.
Priscilla (24:40):
So I think that when I approach it, I think, ok, ai
is a tool that I can use.
I'm not going to use it as theonly way that Buzzsprout
connects with a customer, but itmight be that sometimes, or the
first time that someone reachesout, that's how they connect,
but then they know that I amhere on the other side of
(25:01):
whatever email or phone call andavailable if they need me, and
I feel very strongly that whenAI is used, there should always
be the ability to get to a realperson.
Conor (25:13):
I entirely agree it should be very clear.
Priscilla (25:16):
I don't want any roadblocks that are going to say
OK, you have to sayrepresentative six times before
you get to a representative, orwhatever it is as a company or
as a product Buzzsprout.
I want to make sure that no onefeels like man.
If I'm going to reach out tothe support team, I know I'm not
going to get a person untilI've been on a chat bot for 20
(25:36):
minutes.
And so I want to make sure thatthat human interaction is
available, but not everyone willwant that.
Some people will say, hey, Iactually want, like, immediate
responses and immediate answers,and I don't care if it's an AI,
because what I'm trying to dois log in and I don't want to
wait for Sally to answer theother email that she's working
on before she helps me log in.
I want to be able to log ineven if it's AI, or that's
(25:58):
better than having Sally, and soI think that's where I kind of
look at it is.
There's going to be levelswhere AI can really be a really
great tool to use, and thenthere are going to be levels
that are still going to be thosehuman interactions.
And there are going to belevels that are still going to
be those human interactions andAI can help you make those
stronger because it's takingaway the easier, more
(26:19):
time-consuming tasks that youmight be putting your time in
when you should be buildingrelationships.
Conor (26:24):
Yes.
Priscilla (26:28):
The thing that I love about customer support is the
human interaction side.
So what I want AI to do is comedo the things that keep me from
the human interaction side, andI think that if you approach it
in that way, then those thingscan coexist.
Do you want to shout out anyspecific companies that you
think do bring AI into theircustomer support?
Conor (26:45):
really well, I'll shout out Zapier first.
Jordan (26:48):
Okay.
Conor (26:49):
Because one of my like maxims, one of my philosophies,
is like the best customersupport is not needing customer
support.
Yes, by just having like aproduct that helps you.
And zapier has like greatai-based troubleshooting in
their product.
Now, okay, so if a step breaks,there's just like a
troubleshoot this button, and soit will read the error,
(27:10):
interpret the error and then itwill suggest solutions for it.
And it's just one of those smallthings where I don't need to
spend as much time at all fixingthe thing.
So shout out Zapier for thatthing.
Jordan (27:22):
Yeah, and I love that Zapier has kind of upgraded the
clippy of Microsoft Word.
Conor (27:28):
Yeah.
Jordan (27:29):
In that you run into a thing and it's like uh-oh, you
need help.
And it's like, yes, you needhelp.
It's like, yes, you're back,great, I do.
Conor (27:34):
It looks like you're trying to connect notion to
todoist.
Can I help with that?
Yes, you can.
Thanks, and I will say uh, Ithink intercoms finn is actually
quite useful and I'm sort ofspecializing on implementing and
optimizing intercom forcustomer support at the moment
and I've spent more time liketalking to finn to get answers
(27:56):
than like searching theknowledge base or anything like
that, because it's just so mucheasier to have finn summarize
the answer than me to have to gooff and search it and like
click through a bunch of linksthat's so cool yeah it's funny,
like that's what I want reallyis like I don't want to have to
go and search the knowledge base.
I think the knowledge bases arereally useful and I will often.
Then, after I've got thesummary, I will go to the page,
(28:18):
sometimes to verify theinformation just because it
sounds a bit fishy, or sometimesjust to get a bit more detail
on it.
But yeah, to zapier andintercom as well.
For intercom, it's also thefact that they're like very
openly talking about from acustomer support team's
perspective what it's like touse AI and what they're building
and they're very specificallynot trying to like fire 90% of
(28:41):
their support team.
They're taking this as a way toadd value, via their support
team, to their customers byhaving much faster response
times, by having much moredetailed customer support
knowledge and detailed productknowledge for their support reps
and by specializing them alittle bit more as well.
Priscilla (28:59):
I think it's good to know the companies that are
doing it well so that, as you'retrying to find ways to
integrate this into your dailyworkflow, you know who to look
to, who's already doing itreally well.
What do you think about beinghonest and transparent, about
talking to an AI, like, do yourecommend that people be very
honest or do you say, hey, thatthey don't need to know that
(29:21):
it's an AI?
What do you think?
Conor (29:22):
I think they probably should know.
That's an interesting one,though, because I think the
research is that people havebetter interactions if they
don't know it's an AI, but Istill think that the honesty
aspect is better there, justbecause it's more clear and more
your expectations are setdifferently that when I'm
talking to an AI, I do interactin a different way.
(29:43):
When I'm talking to ChatGPT, Iam asking questions in a
specific way.
When I'm talking to Finn, I'masking questions in a specific
way.
When I'm talking to any kind ofAI, I'm talking to it in a
specific way.
When I talk to a human, I talkin a very different way.
Jordan (29:57):
Yeah, yes.
Conor (29:58):
I am one of those strange people who says please and
thank you a lot of the time tothe AIs, Me too yeah right when
the AIs take over.
Priscilla (30:06):
I want them to know that I was kind to them.
Conor (30:09):
Yes, Exactly, exactly.
We do not.
We want to be like third orfourth against the wall.
It'll be fine, just like dropfirst.
Priscilla (30:17):
Yes, I saw an article recently a coworker sent to me
that was basically saying thatwhen you say please and thank
you to the AI, you're actuallycausing a lot more processing.
Yeah, and it's a bad thing, andyou should not say please and
thank you.
Conor (30:37):
And I thought, well, why don't you take that out on your
end?
Yeah, yeah, I think OpenAI hadsome sort of press release where
they said that they have spenttens of millions of dollars on
customers saying please.
Jordan (30:42):
And thank you.
Conor (30:42):
Yeah, but you know that's OK.
They have plenty of money,Don't worry about them.
Priscilla (30:47):
Yeah, they'll be OK.
So, before we wrap up, what Iwant to give some listeners is
some good advice, some good kindof next steps that they can
take.
So what would you recommend?
Either support professionalswho are on small teams or
leaders, what would yourecommend that they do as those
first steps in implementing AIand bringing it in?
(31:09):
How would you start?
Conor (31:10):
I think that the starting point is like understand what
actually sucks about yourcustomer support at the moment.
Because, not all of us havelike 30 minute response times
for messages, for like emails.
Not all of us pick up the phonein 15 seconds.
Not all of us have like oneminute before your chat gets
picked up by a human team member.
(31:30):
I would say the majority ofcompanies providing customer
support are doing it much slowerthan they want to.
That's true.
Priscilla (31:37):
Yeah, I agree.
Conor (31:38):
There's going to be things that you, as a support
leader, know suck and havesucked for so long, and they
predate you.
It's fine.
It's not your fault.
You're working with limitedresources.
No one blames you.
It's time to just identify whatthose top pain points are for
your customers in their supportexperience and find out how you
(32:00):
can best address those.
So it may not be that you havean opportunity or that the most
pressing matter actually may notbe to get like an AI
conversation agent.
Maybe what you've found ismaybe you've got like super
dialed in follow the sunscheduling and actually you've
got it really well staffed, andactually the problem is that
(32:22):
there's inconsistencies with howcustomer support agents are
answering questions.
Well, okay, maybe the best stepas a starting point is like
finding an AI drivenconversation quality tool that
will plug into your existinghelp desk and give feedback to
your support agents on wherethey're not aligned with tone or
voice, where they're givingincorrect information or those
(32:45):
kind of opportunities Like thatmight be.
Your first step is gettingconsistent, high quality support
interactions before you takeany other opportunity.
But you, as the support leader,you're going to be the one who
actually knows what theseproblems are, or you're going to
hire someone who's veryexpensive, like me, to come in
and tell you what all yourproblems are Exactly?
Yeah, so that's what thestarting point is.
(33:06):
Yeah, alongside that, alsothink about, like, what would
you do every day if you hadunlimited resources?
Like maybe you do somethingonce a quarter at the moment and
you would love to do it likeall the time.
I'm thinking, for example, ifyou're someone who does voice of
customer reporting.
Voice of customer reportingtypically takes like days of
(33:28):
energy and effort and you'reonly working with a small data
set.
And you're only working withlike small data set and you're
only working with likepredefined tags.
And if you forgot to updateyour tagging list to like add in
the new feature that you rolledout three weeks ago, well, I
guess we'll look at that nexttime instead.
And if you want to do thatconsistently pre-AI oh, that's a
lot of effort and you're notgoing to do it every week or
(33:50):
every month.
But, now we have tools that willplug in again, plug into your
existing help desk and that willhelp you, as a support leader,
to demonstrate the value ofcustomer support experiences and
demonstrate the value ofcustomer support by bringing the
voice of the customer to therest of the business and saying,
hey, here's what our customersare struggling with and help you
(34:10):
push for prioritization moreeffectively, instead of just
like one Slack message at a timeinto 16 different product
channels that no one reads andno one gets any traction on.
I would say those are the twothings that you should consider
is like what are the things thatsuck the most and what are the
things that you could do thatyou could change from being like
(34:32):
a once in a blue moon thing tobeing an everyday thing and
everyday activity through thesetwo links.
Priscilla (34:37):
Sometimes you know we talk about AI and it's like, oh
, it's going to take over my job, ok, but there are some aspects
you would like it to do for youbecause you don't like doing
them.
We can love our work and havecertain tasks that we really
dread doing, and how can AI makethat a more enjoyable
experience?
Or how can AI take that off ofyour plate, and so I think
(34:57):
that's another way to like.
If you hate scheduling andyou're like man, having to
figure out the schedule everyweek is so annoying, well, maybe
AI can help you with thescheduling part.
Now, if you love the schedulingpart, ok, then you keep doing
that, but maybe AI comes in andhelps with training new people.
You know, whatever it is, Ithink, looking at the things
that you hate to do during theday, the things that you're like
(35:19):
okay, here I have to do thisagain.
Let AI do it for you and seehow that makes you feel, and
maybe they do a pretty good jobof whatever it is, and then you
can start offloading that ontoAI and opening up your hour to
do more of what you were saying.
What AI software do yourecommend?
You've mentioned Intercom andFin, but what else do you
(35:42):
recommend people test out?
Conor (35:44):
I'm all about like addressing the problem with the
right solution.
So it's going to be like what'sthe best tool for the job?
If I say Intercom, there aregoing to be tons of situations
where Intercom doesn't makesense for someone, where it's
just not the best tool for thejob.
Job, I think my recommendationis actually go to somewhere like
Elevate CX or go to somewherelike support driven and chat to
(36:06):
other people who are in similarsituations.
Like, if you're providing a B2BSaaS tool, you're going to have
different requirements than ifyou're providing, like, b2b
e-commerce support and you'regoing to need different tooling
as a result of that anddifferent products as a result
of that.
Priscilla (36:20):
I think that's a great way to approach it Find
people that are in your space,that are doing it, and get their
recommendations.
Okay, to wrap it up, what isone main takeaway that you would
want someone who's listening tothis episode to take to
remember as they go into theirwork for the rest of the week?
Conor (36:36):
In 10 years, most jobs will probably still exist.
Customer support will probablystill exist.
It'll likely look different.
I think in 10 years, customersupport will be more fun to work
in.
I think customer support willbe more fun to engage with as
customers and a lot of that willbe thanks to, like, proper,
(36:58):
decent people introducing AI inan empathetic and engaging way
with high value opportunities.
I think it is going to get abit weird for a while now, but
hopefully less weird in thefuture.
Priscilla (37:10):
Yeah, and I like the idea of customer support
becoming more fun and morevalued.
Conor (37:16):
Yeah.
Priscilla (37:16):
Well, thanks for joining us, Conor.
This was really fun.
I appreciate the care that youbring into the conversation and
it's nice that you come into itwith.
Hey, we have to be intentionalabout how we're using it.
We can't just haphazardly bringit in and hope that it works
out, that you have to beintentional about how you work
it into your workflows and youhave to be intentional about how
you let it interact with yourcustomers, because you want to
(37:37):
retain that trust and so youcan't just let it on the loose
and hope that it works out.
I appreciate you coming on andsharing your insights and, like
you were saying, if anyone needsmore information or they want
some personalized advice fortheir company where they are,
reach out to Conor because hewill help you.
Conor (37:55):
I will.
I would suggest going tocustomersuccesscx.
If you want to get an emailfrom me tomorrow, go to
customersuccesscx.
Slash daily and I say daily,but technically it's week daily
because I don't really want towrite and send an email on the
weekends.
So it's not really a dailyemail, but it is a week daily
email.
So, I send an email to customersupport leaders every Monday
(38:16):
through Friday and it containsone small short little lesson or
idea or thought about how toprovide a better customer
support experience.
But if you don't like emails atall, you can find me on
LinkedIn Just have a search forConor Pendergrast.
Priscilla (38:31):
And we'll have all of Conor's links in the
description so he's easy to getin touch with.
I would highly recommendsigning up for the newsletter.
It really is just a nice littlenugget of an idea and then the
next day builds on that.
I really like that.
A lot of times there's kind ofa cohesion there from day to day
, and so if you work in customersupport, you absolutely should
(38:51):
subscribe to it.
It's time for Support in RealLife our segment where we
discuss real life supportexperiences.
Support in Real Life, oursegment where we discuss real
life support experiences.
So Jordan has a story, I think,for us today.
Jordan (39:05):
Yeah, okay.
So this one comes from a storyI found about Patagonia.
Obviously, patagonia is alreadypretty famous for treating
their customers really well.
So this woman and her husbandgo to a Patagonia store on the
Upper West Side because herhusband needed a new winter coat
.
Her husband finds a coat heloves and they're up at the
register paying.
And the woman kind of jokinglysays to the employee I'll have
(39:29):
to come back soon.
I have a Patagonia sweater Ilove, but you know, it's getting
holes because I've worn it somuch.
And without missing a beat theemployee says oh well, we
actually consider that a productfailure.
And she tells him no, no, no,it's my fault.
I've been wearing it nonstopfor years.
It doesn't own, you don't oweme anything.
But the employee insists thatPatagonia's philosophy is that
(39:54):
if their product doesn't hold uplong term, they see it as their
responsibility, not thecustomer's.
So right there, and then noreceipt, nothing.
He offers to replace hersweater.
And then it gets even better.
She tries on a new sweater inthe same color, falls in love
with it and ends up buying asecond one in a different color
at full price.
And then they arranged for bothsweaters and her husband's coat
(40:14):
to be shipped from a warehouseso they didn't have to carry
everything around New York Likeit just keeps going.
The story keeps getting betterand better and better.
I love it so much.
Priscilla (40:25):
It's just a great example of really fantastic
customer service.
Yeah, you know it's not like asupport question, but I think
what it reminds me of is theimportance of empowering your
employees to make thesedecisions, because if she had to
go and ask her manager, can Ido this?
And then it would lose theimpact.
It might still be a good result, but you wouldn't have the
(40:47):
impact of being right there atthe cash register and saying, oh
well, that's on us, let me getyou a new sweater and let me
ship it to your house.
And the employee has totalability to do that, and they've
been empowered to do that bytheir leadership.
So, Conor, to make this asupport in real life question,
I'm going to pose a question toyou how would you help a company
that is struggling to empowertheir employees to make those
(41:11):
kinds of decisions to delightcustomers?
Conor (41:13):
Yeah, it's tricky.
I mean, probably you'd come backto like why are we looking at
it in the first place?
Like Patagonia has this asreally as a strong philosophy of
theirs and has done from thestart.
But if I'm talking to acustomer about it, it's like
what's the genesis of what we'retrying to do here?
And if, fundamentally, whatwe're trying to do is create a
much better customer experience,then we've got to find ways of
(41:33):
doing a much better customersupport experience, and removing
those roadblocks and gettingmuch faster resolutions to those
problems by empoweringindividual agents is a great way
of doing it.
Now you could put likethresholds in place for it.
You could say like if it'sunder $75, then go for it, if
it's over $250, you've got toget someone else to look at it.
(41:55):
But you could also put it likea buddy check system in there
rather than going through amanager, so that instead of
going off and having to find amanager, you can talk to a
colleague and just say like hey,I just need a buddy check, that
I'm being sensible in therefund that I'm about to give.
Priscilla (42:09):
Does this make sense to?
Conor (42:10):
you as well.
Priscilla (42:10):
Yeah, and it can have a really positive impact on the
agent or the professional too,because if you're the one
getting to make that call, thatthe customer gets this new
sweater, it can be really funfor you to be like I get to make
this call, I want to give thisto you.
It feels very personal, whereasif you are only following a
process that's in place, then itfeels like, well, this is
(42:31):
coming from Patagonia, it's notcoming from me, so it's a little
less of a personal thing.
But if you make it somethingthat I get to make the call on,
then that I get to make the callon, then I can kind of approach
it with a little bit morepersonality or personal touch
because I am the one making thiscall for you.
But I think you're right, Ithink, putting those guidelines
(42:51):
in place, if you're worriedabout it, if you're nervous to
give away that kind of control,then you can put those
guidelines in place.
You can try it out for a littlebit, do some reporting, see how
it works, pull it back if youneed to and kind of figure that
out until you find a good placewhere you feel comfortable
giving that empowerment to youremployees.
If you have any questions or asupport story or situation that
you would like us to discuss orshout out, you can email us at
(43:12):
happytohelpatbuzzsproutcom ortext the show using the text the
show link in our episodedescriptions, and you may hear
us talk about your story ordiscuss your question on a
future episode.
As always, if you liked thisepisode, please share it with
someone who works in customersupport and leave us a review on
Apple Podcasts.
Thank you all for listening.
Now go and make someone's day.
Welcome to Happy to Help.
A podcast about customersupport from the people at
Buzzsprout.
I'm your host, Priscilla Brooke.
Today we're tackling AI andcustomer support.
Ai is all the rage, but how canwe use it in support to make
our work better, provideexcellent experiences and build
strong customer relationships?
Thanks for joining us.
Let's get into it.
(00:21):
We launched this podcast a yearago and I knew when we launched
it we were going to have to talkabout AI.
Yeah, it was going to happen.
It has been such a hot topicrecently and I was always kind
of like oh, I don't know, whatdo I want to say about AI.
I'm still figuring out what Ithink about AI and how to best
use it.
Yeah, on today, who has somereally great insight into how
(00:43):
you can use AI with yourcustomer support teams to really
continue offeringcustomer-focused support and
using AI as a tool to help youdo that.
So joining us today is ConorPendergrast.
Conor is a customer support andretention specialist with over
15 years of experience.
He has worked as a director ofinformation and communication,
(01:05):
spent 10 years with Expens as adirector of information and
communication, spent 10 yearswith Expensify focusing on
customer success and retention,and currently he is the founder
of customersuccesscx, where hehelps CX leaders grow their
careers, empower their teams anddrive meaningful business
impact.
He also writes a wonderfullyhelpful daily newsletter for
support professionals.
Thanks for joining us, Conor.
Conor (01:24):
Thanks, Priscilla.
That was quite the intro.
I'm impressed with myself.
Priscilla (01:29):
You should be.
I'm impressed with you, andeveryone listening should be
impressed.
We always like to start ourepisodes with a lot of
positivity, and so one thing welike to ask our guests is who
has made your day recently?
Conor (01:41):
I'm going a bit left field here because it has
absolutely nothing to do withwork or anything like that I
will shout out two of my son'sfriends' parents.
So I live in Birmingham theBirmingham that's in the United
Kingdom, not the Birminghamthat's in Alabama for anyone
who's in the United States ofAmerica.
Priscilla (01:56):
Thank you for the clarification.
Conor (01:58):
This is going to surprise you.
There are places outside of theUnited States that have the
same name.
There actually was a York, justa York existed, so anyway.
So I live in Birmingham and wehave the Botanical Gardens here
in Birmingham and it's a lovelyplace.
It's got like nice glass housesand flowers and it actually has
a really nice playground, andso my son's friends' parents
(02:19):
were going there and they justthrew it into the little
WhatsApp group and said, hey,we're going going to go here,
and it turned out to be a reallylovely day.
We rocked up at about 10o'clock, which is when it opened
.
It was sunny.
When I say it was sunny, I meanit was like this is england and
it's april and so you don'treally get consistent dry, warm,
sunny spells, but we did forthe whole day and I mentioned
(02:40):
earlier in the green room that Ihave a little bit of sunburn,
so that's how I got sunburned,but it was just it was just, it
was me, it was my kids, it wasmy son's friends and their
parents, and it was just like areally chill day.
And then, after we finished, weall came home.
I went for a short run, cameback I'd abandoned my son again
at their house.
I came back to collect him andthey were sitting out the front
(03:03):
drinking wine.
So then we segued straight intojust sitting out the front
having a nice drink, gettingmore sunburnt, getting some
pizza and then falling into bed,all of us totally exhausted on
the Sunday.
So it has nothing to do withwork, but I will shout out to
Eleni and Christina, who are thetwo parents that's so great.
Priscilla (03:20):
It is such a good reminder to us to you know,
think of other people.
Reach out to them Like younever know what someone else is
doing with their day.
They might be totally free andable to go and have a fun Sunday
afternoon.
So, yeah, I love that.
Conor (03:32):
Exactly.
Jordan (03:33):
And I'll tell you what when you're a parent and you
have another parent supportsystem where they're like yeah,
the kids can play, you can justdrop them off, go do your
errands.
That is so rare and preciousand you just hold on to those
people as tight as you can.
It's such an amazing thing.
Conor (03:54):
Yes, I agree.
Priscilla (03:56):
Thank you for sharing that, Conor.
Before we get into the world ofAI and how we can use that, I
know I gave a lot of yourbackground in the intro AI and
how we can use that.
I know I gave a lot of yourbackground in the intro but for
anyone who isn't familiar withyou, what else is there for
people to learn about?
What it is that you do and thework you do in customer support
and how you've kind of gonethrough that the last 15 years?
Conor (04:16):
I guess my earliest customer support interactions
where I was doing like paid work.
I worked in a shop that soldsheet music.
So when you're a musician, youhave sheets of music and you
read off them.
And let me tell you, dearlisteners, I was neither a
musician nor really like asalesperson, nor, let me just
say, I was not a good fit forthis role.
But thank you, liz.
(04:36):
Liz was a friend of my mum'sand Liz just gave me the job and
I was awful at the job.
I didn't know anything aboutmusic.
People would come in with themost basic questions and I'd
just be like I don't know.
I got to go and ask Liz, but itwas great.
But I would say, looking backat myself at that point, that
was very much like cutting myteeth on being in an environment
that I didn't know very welland didn't do great at it.
(04:59):
But I'm very proud of the workthat I do now.
It was a far stretch from whereI was.
So after that that wasuniversity.
That was the usual thing ofjust like make a couple of quid
on a Friday, spend it all onFriday night, make a couple more
quid on the Saturday and keepsome of it after Saturday night.
That was the goal.
The goal was like always end upin slight profit.
After you hit the pub on Fridayand Saturday night, it was
(05:27):
Dublin.
All right, I am Irish, so whatcan I say?
After that, I worked for asocial care company.
We created a software productin there and I spun it out as a
startup, moved to London, livedthe London startup dream, which
is mostly like getting trains tothe rest of England and having
terrible Wi-Fi on those trainsand really considering what I'm
doing in my life, did that for ayear and left and joined
Expensify.
I initially joined Expensify,funnily enough, as a sales
(05:50):
generalist, because what Irealized from that earlier role,
that startup role, was that myskills as a salesperson were
lacking, so I needed to work outhow to sell, although I have
different philosophy aroundsales now compared to what I was
expecting to want there.
Within a week of joiningExpensify, we all realized that
there was this new team that wewere creating, which was much
(06:11):
more about account management ina long-term relationship
building way, and so I joinedthat team as one of the first
four people on the team, andthen we evolved that over time,
I eventually, after nine and ahalf years at Expensify was the
senior leader there, which isthe sort of like step below
director Left there in February2024.
And since then I've been workingindependently.
(06:32):
I'm a coaching consultant forcustomer support professionals
and leaders.
So if you have, if you providecustomer support, my job is to
help you provide better customersupport.
And I do that, as I said, boththrough coaching and consulting
customer support, and I do that,as I said, both through
coaching and consulting.
Priscilla (06:45):
So since this is an episode about AI, I figured I
would let AI ask a question tokick us off.
Conor (06:50):
Oh God.
Priscilla (06:58):
Love it, yes, so I asked AI what AI would like to
know about you, and this was thequestion they asked what's one
experience you've had, either asa customer or working in
support, that made you thinkthis is why I love what I do?
Conor (07:07):
So I don't think there's a single experience that really
showed me or reminded me likethis is why I love what I do.
But there was a pattern oftypes of customer support
interactions that I really liked.
Even after nine, 10 years atExpensify, even as a senior
leader, I was still managingsome escalated support
interactions.
And it was the pattern of likeone customer hit a really
(07:28):
peculiar bug and like it wouldtalk on my little curiosity and
I'd be like I there's somethingthere, there's some pattern
there where I think there's somethat I can work out what's
going on.
And you'd like dive into thelogging tool and you'd find that
one weird error and you traceit back to the code base and
you'd be like hang on, this filehasn't been changed in six
years.
How is it possible that noother customers have ever hit
(07:51):
this error?
But for some reason it's likeit's just a bug that has existed
for six years and no one hasever hit it.
We've all seen weird upgradeflows or migrations, or you've
added a new piece offunctionality but you've needed
to change stuff in the back endto support that and because of
that sort of ingrained knowledge, you have just had this like
(08:12):
habit and pattern of how yousleuth on things.
So it's so sort of like weird.
One person will hit this in inin a six-year period.
Bugs that I, I love and I findjust endlessly fascinating and
it'll usually be one of thosethings where, like, I'm the only
person who's patient enough tospend an hour on it.
And that's all it usually like.
It doesn't take a huge amountof work most of the time, but
(08:35):
it's just.
It's kind of fun.
I like it.
Those are the fun ones.
Priscilla (08:37):
Yeah, the investigation.
I love that too.
I love the okay.
Something's going on here.
I know there's an answer.
An answer exists.
Conor (08:46):
How do I find it?
Yes, an answer exists to mostquestions.
Priscilla (08:49):
So let's get into it.
Ai and customer support it'sthe hottest topic right now.
Everyone is talking about it incustomer support, outside of
customer support, just if you'reworking in a professional
setting.
Ai has come up, and so I feellike people can hear AI and
start to worry about theslippery slope of what that
means if we make everythingfully automated and fully AI,
(09:11):
and so I want to start byputting it into a really
practical frame.
So, Conor, how do you use AI inyour day-to-day that makes your
life better.
Conor (09:20):
So, as I mentioned, I coach customer support leaders
to help them provide bettercustomer support.
That will typically involvelike an hour long conversation
every two weeks and one of theways that I use AI in that
process is I have a callrecording tool and the call
recording tool records the calland then gives me a transcript
afterwards but also gives me asummary afterwards.
(09:41):
It's that kind of like reallysimple follow-up step that then
I've dropped that into theNotion page that I've created
with the client and so we canboth go back.
So the following time when it'stime for us to speak again the
day before we speak, I willalways go back and I'll look at
the action points that we pulledout of that.
I'll look at the summary.
I'll look at the video if Ineed to, but very rarely.
(10:10):
The other thing I started doingthis week as well, hot off the
presses this week, was I pulledthe transcript and I ran it
through chat, gpt and I said toit, like what priorities should
I have going into the nextconversation with this person?
And it pulled out things and itpointed out a couple of things
like oh, here's what they saidthey wanted to do and here's an
approach that they said theywanted to avoid.
Here's the style that theywanted to take.
And because I havepre-programmed ChatGPT with my
own set of instructions, itdoesn't do like wishy-washy
(10:34):
stuff.
I've told it to be very directwith me.
You probably know the Kim Scottbook, radical Candor I told
ChatGPT in the custominstructions I said use the
Radical Candor approach.
So every time now it sayssomething to me, it says I'm
going to challenge you on thisspecific point.
And it doesn't say like here'sa gentle nudge or anything like
that.
I hate the gentle nudges, I donot like gentle nudges, I just
(10:55):
want clear and direct feedback.
And so it gave me the prepnotes.
It said like here's the thingsthat you might want to consider
bringing up again.
And then I went into theconversation and then I fed the
transcript from the secondcoaching session into it.
I said, ok, did I hiteverything?
I'm pretty sure I hiteverything.
And it pointed out some placeswhere I had failed in my role as
(11:16):
a coach.
It pointed out places where I'ddone really well and like
clearly pushed this person andclearly like directed them to
make progress and make decisionsthat they were not making.
That's one small way in whichI'm using an LLM like an AI tool
.
The other ways I use it.
I do use it in podcast prep alittle bit as well.
So if I want a quick summary of, like, for example, topics that
(11:37):
haven't come up, gaps thatmight be interesting to talk
about other podcasts that aresimilar to podcasts I've been on
before all those kinds of waysthat you can just save a little
bit of time, a little bit ofGoogling.
That would speed it up.
I use it to rewrite things,sometimes to get it closer to my
voice a little bit, or to like,change it from the style that
(12:00):
I've written an email in to aLinkedIn post, for example.
What I will do personally isI'll avoid using AI to create
things itself, but I will leanon it, like I said, to challenge
my ideas and to give me directfeedback instead.
I think that's a useful thingfor me so far.
The funny thing is it's likethis is all very much in the
context of it's useful forsomeone who's a manager,
probably, or a leader.
(12:21):
But, if you're on the frontlines, if you're working
directly with customers, thisisn't how you're going to be
using it, but there are probablylessons you can pick out from
it as well.
Priscilla (12:30):
Yeah for sure.
So, as someone who has been inthe industry for 15 years, how
have you seen AI really changethe way people approach customer
service, especially in the lastfour to five years, or really
like two to three years, as it'sbeen really hitting the
mainstream and becoming used byso many people?
Conor (12:48):
Yeah, I would say even two years ago, I was extremely
skeptical of how effective AIproducts could be in managing
customer interactions, mostlybecause it seemed really basic
and, a lot of the times, like aperson could do it better with
better information and bettercontext, faster really.
And it's really only in the lasttwo years that I've started to
(13:10):
change my mind even morerecently than that, I'd say just
in the last year of seeingactual progress on AI based
tooling for customer supportprofessionals.
And there are real downsides tothat.
Like I think there have beensome really destructive actions
that some companies have takenin the last year fueled by this
idea that ai can do all of theircustomer support flows and it's
utter, utter rubbish and uttergarbage.
(13:31):
Like I think the real valuethat we're starting to see is
that you can use ai as a tool tobring great customer support,
but that will be by allowingyour support team, and your
success team as well, to engagein higher value, higher leverage
activities or just higher valueactivities in general.
We as humans, are creative, weare insightful, we are
(13:54):
interested, we are curious.
We are the ones who areactually creating these
companies and creating theseproducts and services, and we
are going to know them the best,and we should use AI tools to
support us, not vice versa.
It's just a baffling idea to doit the other way around.
Jordan (14:11):
Yeah, I actually saw a story recently where an AI
customer service bot had fired acustomer and they were just
like it's best for you to cancelyour subscription because I
don't see a route that you cantake.
I was just like, wow, it's bestfor you to cancel your
subscription because I don't seea like a route that you can
take.
I was just like, wow, that'srough.
Priscilla (14:30):
It's dangerous.
Yeah, it's interesting becauseif you look at like support
forums or I know we're both insome different support
communities and on LinkedInyou'll see there's this push and
pull of people who are reallyexcited about the tool AI and
how they can use it to maketheir work better.
But then you'll see people whoare nervous and they're anxious
(14:50):
about what that means for thejob they have and the value they
bring to a company, becausethey've seen things like what
you were talking about and justthese massive layoffs because
people feel like the AI robot orchatbots can fully do what
their role is and the value theybring.
So what would you say to thepeople who are in customer
support right now, who are kindof in that worried stage so
(15:12):
maybe they're not quite ascomfortable with AI yet they're
still like I don't know, I don'tlike this, I want to fight
against it what would you say tothem?
Conor (15:19):
Yeah, I think we have a real challenge in that we do
have sort of like a perfectstorm of really like not great
situations happening in a lot ofways, and customer support has
traditionally not been very wellvalued by companies and it's
been really under-resourced andthese are already unstable
positions in a lot of cases.
I think, broadly speaking, yourwork will look dramatically
(15:43):
different, but I think you willstill have a value to bring to a
company, so long as the companyvalues you already.
Priscilla (15:50):
Right, and don't just completely ignore that AI is
there and hope that it goes away, because I don't think that's
happening.
So what customer support tasksdo you think AI is best suited
for?
What would you give to AI andwhat would you say, oh, I don't
want AI to touch that.
Conor (16:05):
Yeah, it's a tricky question.
The first one is like if thereis a reliable association
between a question and an answer, you can probably have AI
engage in that.
That could be like you have asolid knowledge base great, good
work, you set yourself up forsuccess.
You should have had that anyway.
But now there's like a bigincentive to actually have a
great knowledge base.
(16:26):
It's like before now it waslike oh, let's build a help
center.
Yeah, Okay, great, we threw onetogether, It'll work fine.
And now if you have a wellstructured health center, you're
kind of off to the races andyou can do really great work and
get your AI doing that becauseit's got a solid basis of
information and when itcontradicts what you think, it's
going to be pointing out thatyour knowledge base is wrong.
(16:48):
If you don't have a solid baseof any knowledge, then you're
not really going to get muchbenefit from AI, because it's
just going to be making stuff upand you don't want that.
If you can't afford a wronganswer, you probably shouldn't
get AI involved either.
So that might be, for example,for topics around compliance or
(17:13):
topics around legal questions oranything like that, where you
don't want it sort of firing offsome half-thought-out answer or
its best guess of what it is.
You could probably use AI insummarizing and interpreting and
suggesting, but a human expertshould probably be involved in
reviewing and confirming thosesort of things.
And then there's the likesensitive topics as well.
If you have a product that hasa subscription associated with
(17:34):
it, cancellations and discussingcancellations you probably want
that to go to humans a lot ofthe time as well.
A lot of people will be verycomfortable with having AI
interactions.
Just do the whole cancellationflow for them.
If you want to go down theroute of fully understanding the
problems and seeing if you canrescue an interaction, then you
(17:54):
probably want to have that gothrough humans as well.
Priscilla (17:57):
Yeah, and I think the holiday after hours coverage.
I think that's another way thatAI can really be helpful,
especially if you're a smallteam and you're struggling to
provide coverage for what yourcustomers need or struggling to
take time off, letting AI comein and stand in the gap for you
so that you can have a break andletting AI take that.
(18:19):
And, like you were saying just asecond ago, I liked the idea of
not giving AI the work thatcannot be messed up, kind of the
idea that you know, if it'sthat important, it really should
be someone who can sit thereand give it the focus that it
needs, because with AI, youmight not have a quick review of
what happened and so somethingcould slip under, and so I think
(18:42):
both of those are good examplesof when to use AI, and then
some ways where you know reallyyou might want to be human first
in that situation.
Conor (18:51):
Exactly.
There's also opportunities forlike decisions that are really
about your company and yourcompany's culture.
So one topic that came up thisweek with the client was
thinking about like how, whenyou offer premium support, how
are you framing your premiumsupport?
Should it always go to humansupport agents or should it also
go through your AI agent first?
(19:13):
In that particular case, I wasof the opinion that the average
outcome would be better if itgoes through your AI interaction
first and then, if it's notaddressed, then it just gets
quickly escalated and it's aprioritized response at that
point.
But that's going to be veryparticular to the type of
support that you're giving andwhat the expectations are as
well, because if you're doingpaid support some people might
(19:38):
expect that you always getthrough to a person.
But like that means you have tostaff like 24-7, 365.
That's going to be an expensivepremium support option and if
someone's paying a relativelysmall amount per month for that
premium support option, it's notgoing to be human.
First, and it probablyshouldn't be, because then it
would be a lot slower and peoplewould be a lot more frustrated
by that as well.
Priscilla (19:58):
Yeah, that's a good point.
We don't do premium supporthere at Buzz Brown.
It's something that I don'thave a ton of experience with
myself, kind of having tiers tosupport.
But I can imagine that if I waspaying for this premium tier,
that I wouldn't want an AI to befirst.
But at the same time, if it's anight or a weekend, there's got
(20:19):
to be some understanding that,hey, there might not be a person
ready at my beck and call on anevening or a weekend.
Now, maybe during working hoursI always have a quick response
from a human.
But you might find that there'ssome times when that AI first
can still be the initialexpectation.
So one of the things that we doon this, we're saying, hey, we
(20:54):
really care about human customersupport and feel very strongly
that that is what makes youstand out in your competition is
when you have these humaninteractions through customer
support.
So where does AI live in thatworld of really human,
empathetic, personal customersupport?
Conor (21:14):
God.
It's such a curious topicbecause relationships are
actually created in a whole hostof complex ways and everyone
listening today will haverelationships or feel like they
have relationships withcompanies and they don't know
the people who work there.
And it's the strangest thing.
(21:34):
What I'm talking about iseffectively brand loyalty or
maybe, dear listener, parasocialrelationships, like parasocial
relationships, have become thissort of buzzword.
Talking about podcasts and Ilisten to podcasts I listen to
podcasts all day long.
We do too.
It was the London Marathonrecently and I was following a
(21:54):
couple of people I actually know, but a couple of more like
better known athletes.
It was funny.
Afterwards, I was equallyinvested in knowing how Claire
got on, as knowing how Phillygot on, even though Philly is a
YouTuber and like a runningprofessional athlete like she's
a proper professional athletebut like see videos from her
(22:15):
once to twice a week and Ireally want to follow up on her.
It's not from her having arelationship with me, it's very
one sided.
It's entirely based on what Iam feeling and how I am engaging
with her as a creator, and so Ithink, actually, thought
(22:45):
through that lens, you canactually use AI to promote
relationships between yourcustomers and your company and
products and services as well,because it's effectively just
another way of having consistentbrand perceptions and
consistent brand interactions.
Brand perceptions andconsistent brand interactions,
so long as you have set it up ina way that it all acts
naturally together and feelslike it's all part of a single
ecosystem.
I think that you will be able touse AI as a way of building
relationships with customers aswell, which is sort of the
(23:06):
opposite of what I would everhave said before.
I started thinking about thisand I will say AI doesn't have
empathy.
It's not real.
A lot of the writing that AIdoes it winds me up so much.
The sugar sweet kind of overapologetic stuff.
That's not what I want.
As I went back to earlier.
I like clear, direct and kind,but to the point, without being
(23:29):
over sweet.
So yeah, that's my thought, butlike I'd love to get your take
as well.
Priscilla (23:34):
It's hard, right, it's an interesting question,
right, it's the idea that I wantto use this artificial
intelligence to help buildrelationships with my customers.
But this is not a human person.
But I care about the humaninteraction of myself, as a
person who works for andrepresents Buzzsprout, with the
customers who use Buzzsprout.
(23:54):
I think, personally, that thatis a huge part of offering
really remarkable customerservice, and so when you bring
AI into that, you go okay, well,those things can't coexist,
because one is a robot and oneis a human and you can't have
human interactions with a robot.
It just can't happen.
But I think the thing that youknow you were talking about with
(24:17):
the person you follow andwanting to know how they did in
the marathon Well, the reasonyou care about them is because
they have been authentic inshowing you their lives and
bringing you into their lives,so you are the one with the
feeling and the connection tothem.
They don't have a connection toyou, but they have opened up
themselves so that you canconnect with them.
Conor (24:39):
Yeah.
Priscilla (24:40):
So I think that when I approach it, I think, ok, ai
is a tool that I can use.
I'm not going to use it as theonly way that Buzzsprout
connects with a customer, but itmight be that sometimes, or the
first time that someone reachesout, that's how they connect,
but then they know that I amhere on the other side of
(25:01):
whatever email or phone call andavailable if they need me, and
I feel very strongly that whenAI is used, there should always
be the ability to get to a realperson.
Conor (25:13):
I entirely agree it should be very clear.
Priscilla (25:16):
I don't want any roadblocks that are going to say
OK, you have to sayrepresentative six times before
you get to a representative, orwhatever it is as a company or
as a product Buzzsprout.
I want to make sure that no onefeels like man.
If I'm going to reach out tothe support team, I know I'm not
going to get a person untilI've been on a chat bot for 20
(25:36):
minutes.
And so I want to make sure thatthat human interaction is
available, but not everyone willwant that.
Some people will say, hey, Iactually want, like, immediate
responses and immediate answers,and I don't care if it's an AI,
because what I'm trying to dois log in and I don't want to
wait for Sally to answer theother email that she's working
on before she helps me log in.
I want to be able to log ineven if it's AI, or that's
(25:58):
better than having Sally, and soI think that's where I kind of
look at it is.
There's going to be levelswhere AI can really be a really
great tool to use, and thenthere are going to be levels
that are still going to be thosehuman interactions.
And there are going to belevels that are still going to
be those human interactions andAI can help you make those
stronger because it's takingaway the easier, more
(26:19):
time-consuming tasks that youmight be putting your time in
when you should be buildingrelationships.
Conor (26:24):
Yes.
Priscilla (26:28):
The thing that I love about customer support is the
human interaction side.
So what I want AI to do is comedo the things that keep me from
the human interaction side, andI think that if you approach it
in that way, then those thingscan coexist.
Do you want to shout out anyspecific companies that you
think do bring AI into theircustomer support?
Conor (26:45):
really well, I'll shout out Zapier first.
Jordan (26:48):
Okay.
Conor (26:49):
Because one of my like maxims, one of my philosophies,
is like the best customersupport is not needing customer
support.
Yes, by just having like aproduct that helps you.
And zapier has like greatai-based troubleshooting in
their product.
Now, okay, so if a step breaks,there's just like a
troubleshoot this button, and soit will read the error,
(27:10):
interpret the error and then itwill suggest solutions for it.
And it's just one of those smallthings where I don't need to
spend as much time at all fixingthe thing.
So shout out Zapier for thatthing.
Jordan (27:22):
Yeah, and I love that Zapier has kind of upgraded the
clippy of Microsoft Word.
Conor (27:28):
Yeah.
Jordan (27:29):
In that you run into a thing and it's like uh-oh, you
need help.
And it's like, yes, you needhelp.
It's like, yes, you're back,great, I do.
Conor (27:34):
It looks like you're trying to connect notion to
todoist.
Can I help with that?
Yes, you can.
Thanks, and I will say uh, Ithink intercoms finn is actually
quite useful and I'm sort ofspecializing on implementing and
optimizing intercom forcustomer support at the moment
and I've spent more time liketalking to finn to get answers
(27:56):
than like searching theknowledge base or anything like
that, because it's just so mucheasier to have finn summarize
the answer than me to have to gooff and search it and like
click through a bunch of linksthat's so cool yeah it's funny,
like that's what I want reallyis like I don't want to have to
go and search the knowledge base.
I think the knowledge bases arereally useful and I will often.
Then, after I've got thesummary, I will go to the page,
(28:18):
sometimes to verify theinformation just because it
sounds a bit fishy, or sometimesjust to get a bit more detail
on it.
But yeah, to zapier andintercom as well.
For intercom, it's also thefact that they're like very
openly talking about from acustomer support team's
perspective what it's like touse AI and what they're building
and they're very specificallynot trying to like fire 90% of
(28:41):
their support team.
They're taking this as a way toadd value, via their support
team, to their customers byhaving much faster response
times, by having much moredetailed customer support
knowledge and detailed productknowledge for their support reps
and by specializing them alittle bit more as well.
Priscilla (28:59):
I think it's good to know the companies that are
doing it well so that, as you'retrying to find ways to
integrate this into your dailyworkflow, you know who to look
to, who's already doing itreally well.
What do you think about beinghonest and transparent, about
talking to an AI, like, do yourecommend that people be very
honest or do you say, hey, thatthey don't need to know that
(29:21):
it's an AI?
What do you think?
Conor (29:22):
I think they probably should know.
That's an interesting one,though, because I think the
research is that people havebetter interactions if they
don't know it's an AI, but Istill think that the honesty
aspect is better there, justbecause it's more clear and more
your expectations are setdifferently that when I'm
talking to an AI, I do interactin a different way.
(29:43):
When I'm talking to ChatGPT, Iam asking questions in a
specific way.
When I'm talking to Finn, I'masking questions in a specific
way.
When I'm talking to any kind ofAI, I'm talking to it in a
specific way.
When I talk to a human, I talkin a very different way.
Jordan (29:57):
Yeah, yes.
Conor (29:58):
I am one of those strange people who says please and
thank you a lot of the time tothe AIs, Me too yeah right when
the AIs take over.
Priscilla (30:06):
I want them to know that I was kind to them.
Conor (30:09):
Yes, Exactly, exactly.
We do not.
We want to be like third orfourth against the wall.
It'll be fine, just like dropfirst.
Priscilla (30:17):
Yes, I saw an article recently a coworker sent to me
that was basically saying thatwhen you say please and thank
you to the AI, you're actuallycausing a lot more processing.
Yeah, and it's a bad thing, andyou should not say please and
thank you.
Conor (30:37):
And I thought, well, why don't you take that out on your
end?
Yeah, yeah, I think OpenAI hadsome sort of press release where
they said that they have spenttens of millions of dollars on
customers saying please.
Jordan (30:42):
And thank you.
Conor (30:42):
Yeah, but you know that's OK.
They have plenty of money,Don't worry about them.
Priscilla (30:47):
Yeah, they'll be OK.
So, before we wrap up, what Iwant to give some listeners is
some good advice, some good kindof next steps that they can
take.
So what would you recommend?
Either support professionalswho are on small teams or
leaders, what would yourecommend that they do as those
first steps in implementing AIand bringing it in?
(31:09):
How would you start?
Conor (31:10):
I think that the starting point is like understand what
actually sucks about yourcustomer support at the moment.
Because, not all of us havelike 30 minute response times
for messages, for like emails.
Not all of us pick up the phonein 15 seconds.
Not all of us have like oneminute before your chat gets
picked up by a human team member.
(31:30):
I would say the majority ofcompanies providing customer
support are doing it much slowerthan they want to.
That's true.
Priscilla (31:37):
Yeah, I agree.
Conor (31:38):
There's going to be things that you, as a support
leader, know suck and havesucked for so long, and they
predate you.
It's fine.
It's not your fault.
You're working with limitedresources.
No one blames you.
It's time to just identify whatthose top pain points are for
your customers in their supportexperience and find out how you
(32:00):
can best address those.
So it may not be that you havean opportunity or that the most
pressing matter actually may notbe to get like an AI
conversation agent.
Maybe what you've found ismaybe you've got like super
dialed in follow the sunscheduling and actually you've
got it really well staffed, andactually the problem is that
(32:22):
there's inconsistencies with howcustomer support agents are
answering questions.
Well, okay, maybe the best stepas a starting point is like
finding an AI drivenconversation quality tool that
will plug into your existinghelp desk and give feedback to
your support agents on wherethey're not aligned with tone or
voice, where they're givingincorrect information or those
(32:45):
kind of opportunities Like thatmight be.
Your first step is gettingconsistent, high quality support
interactions before you takeany other opportunity.
But you, as the support leader,you're going to be the one who
actually knows what theseproblems are, or you're going to
hire someone who's veryexpensive, like me, to come in
and tell you what all yourproblems are Exactly?
Yeah, so that's what thestarting point is.
(33:06):
Yeah, alongside that, alsothink about, like, what would
you do every day if you hadunlimited resources?
Like maybe you do somethingonce a quarter at the moment and
you would love to do it likeall the time.
I'm thinking, for example, ifyou're someone who does voice of
customer reporting.
Voice of customer reportingtypically takes like days of
(33:28):
energy and effort and you'reonly working with a small data
set.
And you're only working withlike small data set and you're
only working with likepredefined tags.
And if you forgot to updateyour tagging list to like add in
the new feature that you rolledout three weeks ago, well, I
guess we'll look at that nexttime instead.
And if you want to do thatconsistently pre-AI oh, that's a
lot of effort and you're notgoing to do it every week or
(33:50):
every month.
But, now we have tools that willplug in again, plug into your
existing help desk and that willhelp you, as a support leader,
to demonstrate the value ofcustomer support experiences and
demonstrate the value ofcustomer support by bringing the
voice of the customer to therest of the business and saying,
hey, here's what our customersare struggling with and help you
(34:10):
push for prioritization moreeffectively, instead of just
like one Slack message at a timeinto 16 different product
channels that no one reads andno one gets any traction on.
I would say those are the twothings that you should consider
is like what are the things thatsuck the most and what are the
things that you could do thatyou could change from being like
(34:32):
a once in a blue moon thing tobeing an everyday thing and
everyday activity through thesetwo links.
Priscilla (34:37):
Sometimes you know we talk about AI and it's like, oh
, it's going to take over my job, ok, but there are some aspects
you would like it to do for youbecause you don't like doing
them.
We can love our work and havecertain tasks that we really
dread doing, and how can AI makethat a more enjoyable
experience?
Or how can AI take that off ofyour plate, and so I think
(34:57):
that's another way to like.
If you hate scheduling andyou're like man, having to
figure out the schedule everyweek is so annoying, well, maybe
AI can help you with thescheduling part.
Now, if you love the schedulingpart, ok, then you keep doing
that, but maybe AI comes in andhelps with training new people.
You know, whatever it is, Ithink, looking at the things
that you hate to do during theday, the things that you're like
(35:19):
okay, here I have to do thisagain.
Let AI do it for you and seehow that makes you feel, and
maybe they do a pretty good jobof whatever it is, and then you
can start offloading that ontoAI and opening up your hour to
do more of what you were saying.
What AI software do yourecommend?
You've mentioned Intercom andFin, but what else do you
(35:42):
recommend people test out?
Conor (35:44):
I'm all about like addressing the problem with the
right solution.
So it's going to be like what'sthe best tool for the job?
If I say Intercom, there aregoing to be tons of situations
where Intercom doesn't makesense for someone, where it's
just not the best tool for thejob.
Job, I think my recommendationis actually go to somewhere like
Elevate CX or go to somewherelike support driven and chat to
(36:06):
other people who are in similarsituations.
Like, if you're providing a B2BSaaS tool, you're going to have
different requirements than ifyou're providing, like, b2b
e-commerce support and you'regoing to need different tooling
as a result of that anddifferent products as a result
of that.
Priscilla (36:20):
I think that's a great way to approach it Find
people that are in your space,that are doing it, and get their
recommendations.
Okay, to wrap it up, what isone main takeaway that you would
want someone who's listening tothis episode to take to
remember as they go into theirwork for the rest of the week?
Conor (36:36):
In 10 years, most jobs will probably still exist.
Customer support will probablystill exist.
It'll likely look different.
I think in 10 years, customersupport will be more fun to work
in.
I think customer support willbe more fun to engage with as
customers and a lot of that willbe thanks to, like, proper,
(36:58):
decent people introducing AI inan empathetic and engaging way
with high value opportunities.
I think it is going to get abit weird for a while now, but
hopefully less weird in thefuture.
Priscilla (37:10):
Yeah, and I like the idea of customer support
becoming more fun and morevalued.
Conor (37:16):
Yeah.
Priscilla (37:16):
Well, thanks for joining us, Conor.
This was really fun.
I appreciate the care that youbring into the conversation and
it's nice that you come into itwith.
Hey, we have to be intentionalabout how we're using it.
We can't just haphazardly bringit in and hope that it works
out, that you have to beintentional about how you work
it into your workflows and youhave to be intentional about how
you let it interact with yourcustomers, because you want to
(37:37):
retain that trust and so youcan't just let it on the loose
and hope that it works out.
I appreciate you coming on andsharing your insights and, like
you were saying, if anyone needsmore information or they want
some personalized advice fortheir company where they are,
reach out to Conor because hewill help you.
Conor (37:55):
I will.
I would suggest going tocustomersuccesscx.
If you want to get an emailfrom me tomorrow, go to
customersuccesscx.
Slash daily and I say daily,but technically it's week daily
because I don't really want towrite and send an email on the
weekends.
So it's not really a dailyemail, but it is a week daily
email.
So, I send an email to customersupport leaders every Monday
(38:16):
through Friday and it containsone small short little lesson or
idea or thought about how toprovide a better customer
support experience.
But if you don't like emails atall, you can find me on
LinkedIn Just have a search forConor Pendergrast.
Priscilla (38:31):
And we'll have all of Conor's links in the
description so he's easy to getin touch with.
I would highly recommendsigning up for the newsletter.
It really is just a nice littlenugget of an idea and then the
next day builds on that.
I really like that.
A lot of times there's kind ofa cohesion there from day to day
, and so if you work in customersupport, you absolutely should
(38:51):
subscribe to it.
It's time for Support in RealLife our segment where we
discuss real life supportexperiences.
Support in Real Life, oursegment where we discuss real
life support experiences.
So Jordan has a story, I think,for us today.
Jordan (39:05):
Yeah, okay.
So this one comes from a storyI found about Patagonia.
Obviously, patagonia is alreadypretty famous for treating
their customers really well.
So this woman and her husbandgo to a Patagonia store on the
Upper West Side because herhusband needed a new winter coat
.
Her husband finds a coat heloves and they're up at the
register paying.
And the woman kind of jokinglysays to the employee I'll have
(39:29):
to come back soon.
I have a Patagonia sweater Ilove, but you know, it's getting
holes because I've worn it somuch.
And without missing a beat theemployee says oh well, we
actually consider that a productfailure.
And she tells him no, no, no,it's my fault.
I've been wearing it nonstopfor years.
It doesn't own, you don't oweme anything.
But the employee insists thatPatagonia's philosophy is that
(39:54):
if their product doesn't hold uplong term, they see it as their
responsibility, not thecustomer's.
So right there, and then noreceipt, nothing.
He offers to replace hersweater.
And then it gets even better.
She tries on a new sweater inthe same color, falls in love
with it and ends up buying asecond one in a different color
at full price.
And then they arranged for bothsweaters and her husband's coat
(40:14):
to be shipped from a warehouseso they didn't have to carry
everything around New York Likeit just keeps going.
The story keeps getting betterand better and better.
I love it so much.
Priscilla (40:25):
It's just a great example of really fantastic
customer service.
Yeah, you know it's not like asupport question, but I think
what it reminds me of is theimportance of empowering your
employees to make thesedecisions, because if she had to
go and ask her manager, can Ido this?
And then it would lose theimpact.
It might still be a good result, but you wouldn't have the
(40:47):
impact of being right there atthe cash register and saying, oh
well, that's on us, let me getyou a new sweater and let me
ship it to your house.
And the employee has totalability to do that, and they've
been empowered to do that bytheir leadership.
So, Conor, to make this asupport in real life question,
I'm going to pose a question toyou how would you help a company
that is struggling to empowertheir employees to make those
(41:11):
kinds of decisions to delightcustomers?
Conor (41:13):
Yeah, it's tricky.
I mean, probably you'd come backto like why are we looking at
it in the first place?
Like Patagonia has this asreally as a strong philosophy of
theirs and has done from thestart.
But if I'm talking to acustomer about it, it's like
what's the genesis of what we'retrying to do here?
And if, fundamentally, whatwe're trying to do is create a
much better customer experience,then we've got to find ways of
(41:33):
doing a much better customersupport experience, and removing
those roadblocks and gettingmuch faster resolutions to those
problems by empoweringindividual agents is a great way
of doing it.
Now you could put likethresholds in place for it.
You could say like if it'sunder $75, then go for it, if
it's over $250, you've got toget someone else to look at it.
(41:55):
But you could also put it likea buddy check system in there
rather than going through amanager, so that instead of
going off and having to find amanager, you can talk to a
colleague and just say like hey,I just need a buddy check, that
I'm being sensible in therefund that I'm about to give.
Priscilla (42:09):
Does this make sense to?
Conor (42:10):
you as well.
Priscilla (42:10):
Yeah, and it can have a really positive impact on the
agent or the professional too,because if you're the one
getting to make that call, thatthe customer gets this new
sweater, it can be really funfor you to be like I get to make
this call, I want to give thisto you.
It feels very personal, whereasif you are only following a
process that's in place, then itfeels like, well, this is
(42:31):
coming from Patagonia, it's notcoming from me, so it's a little
less of a personal thing.
But if you make it somethingthat I get to make the call on,
then that I get to make the callon, then I can kind of approach
it with a little bit morepersonality or personal touch
because I am the one making thiscall for you.
But I think you're right, Ithink, putting those guidelines
(42:51):
in place, if you're worriedabout it, if you're nervous to
give away that kind of control,then you can put those
guidelines in place.
You can try it out for a littlebit, do some reporting, see how
it works, pull it back if youneed to and kind of figure that
out until you find a good placewhere you feel comfortable
giving that empowerment to youremployees.
If you have any questions or asupport story or situation that
you would like us to discuss orshout out, you can email us at
(43:12):
happytohelpatbuzzsproutcom ortext the show using the text the
show link in our episodedescriptions, and you may hear
us talk about your story ordiscuss your question on a
future episode.
As always, if you liked thisepisode, please share it with
someone who works in customersupport and leave us a review on
Apple Podcasts.
Thank you all for listening.
Now go and make someone's day.
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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.