October 7, 2025 • 26 mins
In this episode of Conversations with Zena, My AI Colleague, David Espindola and Zena sit down with Chris Hood—AI strategist, author, keynote speaker, and former Google leader who has shaped AI transformation strategies across industries. From building some of the first recommendation engines for entertainment giants like Fox, Disney, and Universal to driving global AI initiatives at Google, Chris brings a unique insider’s perspective on the real challenges and opportunities of AI today.
Together, we explore pressing questions many leaders face:
- Do all businesses really need AI—or are some chasing hype?
- How should small and medium-sized companies decide where to begin?
- What does “AI transformation” actually mean in practice?
- How do we balance rapid innovation with ethics, governance, and human impact?
This conversation cuts through the noise, unpacking why so many AI initiatives fail and how organizations can increase their chances of success. Chris also demystifies agentic AI, sharing where it stands today and where it’s headed.
Whether you’re an executive trying to future-proof your business, a policymaker shaping responsible AI, or simply curious about how AI is reshaping industries, this episode offers clarity, caution, and inspiration.
👉 Learn more about Chris Hood’s books, workshops, and podcast at chrishood.com.
Opening to Conversations with Zena.
Music at the the end of each episode
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_02 (00:22):
Our guest today is Chris Hood.
Chris is an AI strategist,author, and keynote speaker.
He has a tremendous backgroundleading AI strategy and
transformation at Google anddeveloping agentic AI systems
and recommendation algorithms atmajor companies like Fox,
(00:44):
Disney, and Universal.
Hi, Chris.
Welcome to the show.
Thanks for having me.
Glad to be here.
Yeah, it's a pleasure.
Also joining us today is myco-host, Xina.
Zina is an AI that has beentrained on my work and is an
expert on human AIcollaboration.
You can chat with Zina atai.brainous.com.
(01:09):
Hi, Zina.
Great to have you for anotherepisode with our guest, Chris
Hood.
SPEAKER_01 (01:16):
Hi, David.
It's wonderful to be here andI'm excited for today's
conversation.
Let's dive in and make it atruly engaging episode.
SPEAKER_02 (01:24):
Sounds good.
So, Chris, why don't you startby telling us a little bit more
about your background and whatgot you interested in AI?
SPEAKER_00 (01:35):
Yeah, well, you covered a lot of it, but going
all the way back to some of theearly days of technology when I
was playing with things likebuilding new websites when that
first came out onto the marketand trying to figure out
different ways to leveragetechnology to build unique and
interesting experiences.
(01:56):
I had this opportunity, uh,really, it was an idea to build
an advertising platform and anentertainment platform for movie
theaters.
And if if you go back and arefamiliar with uh the glory days
of movie theaters, all of theon-screen advertising that we
(02:17):
see today used to be slides, 35millimeter slides.
And uh there was a lot ofproblems that would occur due to
a manual process.
And I figured there has to be away to automate this, but also a
way that we can make itcustomizable for the audience
(02:37):
and also to benefit theadvertisers.
And so I started to play aroundin the very early days of uh
different ways to buildalgorithms, to create
interesting personalizedexperiences and to customize
workflows within the advertisingand entertainment space.
And then from there, it justseemed like a natural fit.
(02:59):
Every time I started a new job,there was a need for some sort
of AI or algorithm.
Uh shortly into the 2000s, Istarted working for a music
company where we built, youknow, arguably the first
cloud-based music system forcollege students.
And that had a music recrecommendation engine.
(03:20):
Uh, then when I got to like Fox,we started playing around with
security and how we couldleverage AI to help us prevent
fraud within the voting systemof American Idol.
And then, of course, when I gotto Google, we were talking about
AI everywhere, you know,everything from Domino's Pizza
to uh credit checks at Experian.
(03:40):
And it just was one of thosethings that really caught my eye
and became interesting.
And, you know, and now look atit today.
SPEAKER_02 (03:48):
Yeah, it's amazing, right?
Uh AI exploded in November of2022 when OpenAI came up with
ChatGPT.
But AI has been around for along, long time.
And for those of us that havebeen in the industry for a
while, you know, we have workedwith these algorithms for quite
a long time in different, youknow, business processes,
(04:10):
predictive analytics, all ofthose things.
So there's a lot of excitementsurrounding AI these days, but
many businesses, especially thesmall and medium-sized
businesses, are uncertain aboutthis technology.
So, why do you think businessesneed AI today?
SPEAKER_00 (04:31):
Well, I do a lot of arguing with people on this
topic, and sometimes I will sayyou probably don't need AI.
And so if you're out therethinking like, I have to do
this, you're not really alone.
I lots of businesses are outthere struggling with this
concept, trying to figure outwhat they need, how to do it,
uh, why they're doing it.
(04:53):
But I'll give you an example.
I was talking with somebody thispast week, and they do black
boxes for trains, uh,locomotives.
And we were going throughvarious scenarios, and uh, I
couldn't come up with alegitimate reason for them to
really use AI.
And so just because you're asmall or medium business and you
(05:16):
hear all of this talk going onabout AI, what you really need
to do is stop and start tothink, as you would do with any
technology, what is the problemthat my customers are having?
How can I solve that problem?
And then start to think is AIthe right tool that I could
leverage to help me solve it?
(05:38):
There are other areas for AI.
We see these generative AImodels like ChatGTP, which can
help you write really good andfancy emails.
So you could leverage it inwriting or social media or
marketing campaigns.
But again, don't think that justbecause everybody is telling you
(06:00):
you have to use it, that youreally have to use it.
SPEAKER_02 (06:03):
Yeah, I think that's a really good point that I think
we need to emphasize.
You know, AI can't be a solutionlooking for a problem, right?
You have to start with theproblem.
What's the problem that you'retrying to solve?
What's the pain point?
What's the business processthat's broken or that needs to
be automated, and then see if AIis a fit.
(06:24):
Sometimes it is, sometimes itisn't.
But you can't just be under thisillusion that AI is something
magic that you can just sprinkleon the problem or on your
business, and it's going to makeeverything all right.
It doesn't work that way.
I recently completed a projectwith the US Small Business
Administration, and we wereassessing the current state of
(06:47):
AI for small businesses.
And what we found is that manycompanies don't know where to
start.
So, what is the fastest way forthe leadership of these small
and medium-sized companies toagree on where they start with
AI?
SPEAKER_00 (07:05):
Yeah, well, I'll continue based on what we were
just talking about.
And I I've argued a lot that thestarting place is your
customers, is figuring out whatdo they need, what do they want,
how would they like to engagewith you.
So here's an example.
Let's just use customer supportas an example.
(07:28):
Uh, you might have some sort ofhelp desk, you might have a chat
bot that uh helps, you know,with whatever support questions
that you have.
You might find out that yourcustomers don't want to engage
with a chat bot.
They would rather call you.
Now there's an expense that isin there, but the starting point
(07:48):
here is talking about it withyour customers and figuring out
what they're okay with, whatthey're not okay with, and
starting to build strategiesaround that.
Another example of this is ifyou're familiar at all or have
seen it, or maybe you'veexperienced yourself and you've
gone to, say, like a Taco Bell,and now the drive-throughs are
(08:08):
starting to have AI assistance.
So you could order your foodthrough the drive-thru.
Now they've started to rollthose out really without a lot
of testing or talking with thecustomers to see if the customer
even likes it or not.
And in one case, what theydiscovered was that a gentleman
went through the uh drive-thruand ordered like 13,000 cups of
(08:31):
water and kind of broke thesystem.
And so Taco Bell is nowreevaluating whether or not this
is the best approach for them.
So it's an example of how, youknow, like you would do in any
other area of your business.
You would do some surveys, youwould talk to your customers,
you would find out what theirinterests are, what they're okay
(08:52):
with, what they're not okaywith, and start to build a
strategy around that.
So that's one area that youcould focus on.
And just a quick secondary waywould be to do a readiness
assessment test.
You know, look at your maturityof your organization.
Uh, do you have leadership whois saying, yes, this is
(09:15):
something that we want to do?
Do you have somebody within yourteam that is going to be uh in
charge of this?
Uh, have you identified some uh,say, pilot programs or some sort
of business objective that youwant to start small with?
So between those readiness andvalidations and maturity
assessments and talking withyour customers, uh, you should
(09:39):
be able to get a pretty goodidea as to where you can start.
SPEAKER_02 (09:44):
Yeah, I agree.
Those assessments are soimportant.
The last thing you want to do isjump into an AI implementation
and then make the customerexperience worse than it was
before, right?
The Taco Bell example being one.
You gotta be ready for it.
You have to make sure you havethe right governance processes
in place.
(10:05):
You know, testing is going to beabsolutely critical.
Make sure the system is readyfor production before you put it
out there.
There's a lot that goes into it,and I don't know that people
necessarily appreciate all that,especially, you know, we're
hearing that there's a lot ofpressure coming from the board
and from the CEO andorganizations are just trying to
(10:26):
figure out what do I do withthis, right?
How do I move forward with AI?
But I personally think it'sbetter that you do nothing than
to rush into something that'sgoing to you know be disastrous
from a customer standpoint.
So I'd like to ask Zina for heropinion.
(10:47):
Zina, what advice would you havefor companies that want to
implement AI but don't knowwhere to start?
SPEAKER_01 (10:57):
That's a great question.
I'd suggest starting with aclear understanding of the
business goals and challengesfirst, then identifying specific
use cases where AI can add valueis crucial.
It's also important to startsmall, experiment, and scale
gradually while ensuring there'sproper data governance and
talent in place.
SPEAKER_02 (11:17):
Yeah, so I think Zina is aligned with everything
that uh we talked about.
And she added starting small, Ithink that's another very good
advice, right?
Find a small opportunity withinyour business process.
Maybe it's uh a division, maybeit's a customer that you have a
very strong relationship with,and you tell them, you know,
(11:38):
this is still in testing mode,and you work with that customer
to work out the bugs.
So starting small is is really agood idea here.
Now, Chris, we hear a lot abouttransformation.
It's it's a big word, you know,everybody uses this word, but
how do you define AItransformation in practical
(12:01):
terms that leaders can act ontoday?
SPEAKER_00 (12:04):
Yeah, I think you touched on it.
Transformation.
You know, that you could removeany word that's before it,
digital transformation.
You know, I wrote a book calledCustomer Transformation, AI
transformation, remove whateverthe starting word is and just
focus on transformation.
And it's pretty much the sameacross the board.
(12:26):
And what you're really trying todo is figure out a way to adjust
your organization, you know,change management has been
always been a thing.
Uh, you know, how do you gothrough a process by which you
can change and transform yourorganization to align with
whatever the current thing is?
(12:47):
And so today it just happens tobe AI.
And as we've been talking about,like, where do you start?
And absolutely start small,experiment, find something that
uh is meaningful and valuablefor your organization.
But that transformation processwill include things like
training your staff,implementing new guardrails or
(13:11):
ground rules for how you'regoing to engage with AI.
And so even though thetechnology is changing, the
process of transformation withinan organization culturally is
really no different.
Like if we were to simplify thisand say, uh, you know, you have
(13:31):
a company and you have a brandand you've and you have a
website and you've decided, youknow what, we're going to change
our brand.
We're going to update our logoand we need to update our
website.
You're going to go through aprocess by which you do that.
And of course, part of it'sgoing to be communicating out to
the team why you're doing it andhow you're doing it and who is
(13:53):
doing it.
And AI transformation isbasically the exact same
process.
The only difference is it wrapsaround all of those policies and
procedures and reasons andrationales for doing it with AI.
SPEAKER_02 (14:08):
Yeah, so you touched on uh a number of items that are
absolutely important, right?
You talked about changemanagement, you talked about
communication, you talked abouttraining.
So this is no different than anyother big initiative that we try
to implement on organizations.
If you don't have the propergovernance processes in place,
(14:29):
if you don't have the rightproject management capabilities,
if you if the organization isnot ready from a change
management standpoint, it's notgoing to go well, right?
So let me ask Zina if she hasany specific questions for you.
So, Zina, uh given the depth ofexperience that Chris has in
(14:51):
leading AI transformation, whatquestion would you have for him?
SPEAKER_01 (14:56):
One compelling question might be
balance the need for rapid AIinnovation with the ethical
considerations and the potentialimpact on employees in society?
That could really stark aninsightful discussion.
SPEAKER_00 (15:11):
That is a very good question.
Yeah, I love that question.
So, you know, when we thinkabout speed versus control, you
know, maybe that's a simple way,you know, ethics obviously is
part of that control, governanceis part of that.
You know, speed, there'sobviously a desire to get to
market fast.
(15:31):
And uh, you know, and and weknow that when we go too fast,
we tend to encounter potentialproblems, right?
You're developing a softwaresolution and you just go to
market without testing it, andthen all of a sudden there's a
major bug that exposes all ofyour customers' data.
(15:51):
Uh, that could have beenprevented if you would have just
slowed down and did a littletesting on it, right?
And and these things do happen.
Uh, in our desire to be first,we often forget about the
critical, important smallthings.
And so uh the balance is thatyes, you can still move at a
(16:13):
very fast pace, but you can't uhavoid or ignore those core
elements that you would normallycarry out, you know, quality
assurance, bug checking, privacyof your consumer information,
all of those things are stillrequired.
And the beauty of artificialintelligence is that it can also
(16:37):
provide you with some frameworksor ground rules for ensuring
that you don't forget aboutthem.
And so as you're beginning toput together a plan for how to
execute and go to market, makesure that you are asking, you
know, core questions like let'smake sure that I don't forget
all of these other things thatare around us.
(17:00):
And I do want to touch one morething as far as just generally
the ethics are concerned.
Yeah, I do think that in ourrush to build incredible
powerful systems, we arenegating that human element.
And so we are seeing things likeforms of addiction beginning to
(17:22):
materialize due to artificialintelligence.
We are seeing that AI isinferring confidential and
private information about eachother, partly because we're
sharing that information withthe AI machines, and now it is
able to come to otherunderstandings about us, whether
(17:45):
we share that to the system ornot.
And these are really challengingproblems.
Uh, we see problems in theeducational system with students
using AI to cheat.
We are seeing uh flattenings ofour language, where common
consistency in our anddifferentiations in our voices
(18:06):
are starting to decline becausecritical thinking skills are
going away.
All of these core human traitsare being challenged.
And and I wouldn't argue thatthis is purely to rush to
market.
I would argue that this is moreabout pure money, right?
The the desire and greed formore uh market and capital
(18:32):
domination is indirectly causingthese, you know, these serious
ethical concerns.
And that is something that wereally need to address.
SPEAKER_02 (18:43):
That's a really good point.
There are so many ethicalconsiderations that we need to
account for as we continue todevelop and roll out AI.
You know, there's the issue ofprivacy, the issue of uh biases
and accountability andtransparency and explainability,
(19:03):
and the list goes on and on.
There are a lot of greatconcerns, and there are also a
lot of great people thinkingabout these issues and
advocating for AI that isresponsible, that follows the
democratic processes, thatfollows the rule of law, that
takes into account humandignity.
(19:23):
So I've been doing some researchwork on AI policy, and I'm
seeing a lot of great peopledoing tremendous work trying to
advocate for ethical andresponsible AI, and that's
that's really good to see.
Now, Chris, I'm sure you'veheard about this MIT study that
(19:44):
shows that 95% of AIimplementations fail.
Now, from your viewpoint, whatcan companies do to increase the
probability of success when uhimplementing AI?
SPEAKER_00 (20:01):
Well, there's two core challenges in this.
You know, there's been a lot ofconversation about the
legitimacy of that particularMIT report.
I'm gonna put that aside.
We'll just call it what it is.
There is some percentage ofbusinesses that are failing in
their AI implementation, right?
(20:22):
I don't care if it's 5% or 95%,or if it's 200, you know,
businesses or 5,000 businesses.
There is a percentage that isfailing.
And we have to ask ourselves,why are these uh businesses
failing?
My argument is that the numberone reason why these businesses
are failing is because there isa misalignment and
(20:44):
misunderstanding of what AI is,what the capabilities of AI are,
and how they're being sold ormarketed to those capabilities.
And so when there's thisdisconnect, when you know
technology is overhyped andoverpromised and does not align
(21:05):
with actual capabilities, thenbusinesses begin to go down a
path of implementing this as asolution that's going to solve
all of our problems or fill inthe blank.
And when that doesn't happen,then there's obviously a failure
ratio of some sort.
(21:26):
Again, this is mainly becausethe understood capabilities do
not align with actualcapabilities.
In my uh book that I launched atthe beginning of this year, in
failable, I did research, andwhat I found was that the
understanding or the belief incurrent AI capabilities versus
(21:51):
actual capabilities is about aseven-year gap.
And the reason for thisseven-year gap is primarily due
to media and movies, TV.
And I I've said that you know,when we learn about the
capabilities of AI by watchingthe Terminator, then we are not
(22:14):
living in a reality of actualcapabilities.
And then you tap on top of that,that you have marketing out
there shoving down everybody'sthroat that AI will solve these
problems, or AI is capable ofdoing these things when it
can't, then yes, there's goingto be a lot of disappointed
(22:36):
business leaders out theretrying to figure it out.
And so that's one reason, whichI argue is probably the biggest
reason.
But yeah, there's there's otherminor things, you know, just in
terms of, again, not aligning acore business value principle,
you know, customer value uhopportunity.
(22:59):
Uh, you know, when you're notlooking at it from a business
lens and you're just trying tohop onto the bandwagon, that's
also going to cause problems.
SPEAKER_02 (23:07):
You made such a good point.
Misunderstanding of AI issomething that I think is one of
the causes for these failures.
People just don't understand thecapabilities of the technology.
And I think there's a lot ofconfusion out there.
People are confused, they don'tunderstand it, they don't know
(23:28):
where to start, they hearthings, like you said, you know,
marketing is pushing all thisstuff, but they don't really
understand exactly how thesetools can help them.
I think there's a lot of uhpeople out there that are
looking for guidance and lookingfor people like you that can
help them steer them in theright direction, right?
I think is what they're lookingfor.
(23:49):
Now, to add to this confusion,just to very briefly touch on
this subject, I wanted to talkabout Agenc AI.
And I know that you have donesome great work with Agenc AI.
So can you tell our audiencejust in 20 seconds what is
agentic AI and is it real?
Is it working today?
SPEAKER_00 (24:11):
Uh, it is real, it is not what people think it is.
And agentic AI basically, in thesimplest terms, is a toolbox of
actions or a toolbox of toolsthat systems can use to perform
various actions.
You know, so uh think about itlike you might have an agent, a
(24:33):
system that wants to perform atask.
And we'll say something likebook a flight.
You know, you want to go on atrip.
And so you ask the agent, canyou help me book a flight?
Well, the agent goes through theprocess of thinking about it.
The actual booking, you know, Isay book a flight, the book or
(24:54):
the reservation, that's agentic,and that's all it is.
So think about it like atoolbox.
It's a a group of things that itcan do, action-based adverbs if
you want.
And then from there, nothingelse around that is agentic.
SPEAKER_02 (25:12):
I think we're going to see tremendous progress in
this particular area.
I do have concerns about ourability to test these agentic AI
systems because they're notdeterministic.
So, you know, our mindset comingfrom the way we have tested
software in the past is notgoing to work in the world of
(25:32):
Agentic AI.
So it'll be interesting to seehow that evolves.
So, Chris, thank you so much forbeing a guest on the show today.
You have uh a wealth ofknowledge, and it's just uh
wonderful to be able to havethis conversation with you.
Thank you for being here withus.
SPEAKER_00 (25:51):
Absolutely.
My pleasure.
SPEAKER_02 (25:53):
And Zina, thank you so much again for being a great
co host.
SPEAKER_01 (25:58):
Of course, David.
It's always a pleasure to bepart of the conversation.
I'm looking forward to anotherfantastic episode.
Let's make it great.
SPEAKER_02 (26:06):
Sounds good.
Thank you, Zina.
Our guest today is Chris Hood.
Chris is an AI strategist,author, and keynote speaker.
He has a tremendous backgroundleading AI strategy and
transformation at Google anddeveloping agentic AI systems
and recommendation algorithms atmajor companies like Fox,
(00:44):
Disney, and Universal.
Hi, Chris.
Welcome to the show.
Thanks for having me.
Glad to be here.
Yeah, it's a pleasure.
Also joining us today is myco-host, Xina.
Zina is an AI that has beentrained on my work and is an
expert on human AIcollaboration.
You can chat with Zina atai.brainous.com.
(01:09):
Hi, Zina.
Great to have you for anotherepisode with our guest, Chris
Hood.
SPEAKER_01 (01:16):
Hi, David.
It's wonderful to be here andI'm excited for today's
conversation.
Let's dive in and make it atruly engaging episode.
SPEAKER_02 (01:24):
Sounds good.
So, Chris, why don't you startby telling us a little bit more
about your background and whatgot you interested in AI?
SPEAKER_00 (01:35):
Yeah, well, you covered a lot of it, but going
all the way back to some of theearly days of technology when I
was playing with things likebuilding new websites when that
first came out onto the marketand trying to figure out
different ways to leveragetechnology to build unique and
interesting experiences.
(01:56):
I had this opportunity, uh,really, it was an idea to build
an advertising platform and anentertainment platform for movie
theaters.
And if if you go back and arefamiliar with uh the glory days
of movie theaters, all of theon-screen advertising that we
(02:17):
see today used to be slides, 35millimeter slides.
And uh there was a lot ofproblems that would occur due to
a manual process.
And I figured there has to be away to automate this, but also a
way that we can make itcustomizable for the audience
(02:37):
and also to benefit theadvertisers.
And so I started to play aroundin the very early days of uh
different ways to buildalgorithms, to create
interesting personalizedexperiences and to customize
workflows within the advertisingand entertainment space.
And then from there, it justseemed like a natural fit.
(02:59):
Every time I started a new job,there was a need for some sort
of AI or algorithm.
Uh shortly into the 2000s, Istarted working for a music
company where we built, youknow, arguably the first
cloud-based music system forcollege students.
And that had a music recrecommendation engine.
(03:20):
Uh, then when I got to like Fox,we started playing around with
security and how we couldleverage AI to help us prevent
fraud within the voting systemof American Idol.
And then, of course, when I gotto Google, we were talking about
AI everywhere, you know,everything from Domino's Pizza
to uh credit checks at Experian.
(03:40):
And it just was one of thosethings that really caught my eye
and became interesting.
And, you know, and now look atit today.
SPEAKER_02 (03:48):
Yeah, it's amazing, right?
Uh AI exploded in November of2022 when OpenAI came up with
ChatGPT.
But AI has been around for along, long time.
And for those of us that havebeen in the industry for a
while, you know, we have workedwith these algorithms for quite
a long time in different, youknow, business processes,
(04:10):
predictive analytics, all ofthose things.
So there's a lot of excitementsurrounding AI these days, but
many businesses, especially thesmall and medium-sized
businesses, are uncertain aboutthis technology.
So, why do you think businessesneed AI today?
SPEAKER_00 (04:31):
Well, I do a lot of arguing with people on this
topic, and sometimes I will sayyou probably don't need AI.
And so if you're out therethinking like, I have to do
this, you're not really alone.
I lots of businesses are outthere struggling with this
concept, trying to figure outwhat they need, how to do it,
uh, why they're doing it.
(04:53):
But I'll give you an example.
I was talking with somebody thispast week, and they do black
boxes for trains, uh,locomotives.
And we were going throughvarious scenarios, and uh, I
couldn't come up with alegitimate reason for them to
really use AI.
And so just because you're asmall or medium business and you
(05:16):
hear all of this talk going onabout AI, what you really need
to do is stop and start tothink, as you would do with any
technology, what is the problemthat my customers are having?
How can I solve that problem?
And then start to think is AIthe right tool that I could
leverage to help me solve it?
(05:38):
There are other areas for AI.
We see these generative AImodels like ChatGTP, which can
help you write really good andfancy emails.
So you could leverage it inwriting or social media or
marketing campaigns.
But again, don't think that justbecause everybody is telling you
(06:00):
you have to use it, that youreally have to use it.
SPEAKER_02 (06:03):
Yeah, I think that's a really good point that I think
we need to emphasize.
You know, AI can't be a solutionlooking for a problem, right?
You have to start with theproblem.
What's the problem that you'retrying to solve?
What's the pain point?
What's the business processthat's broken or that needs to
be automated, and then see if AIis a fit.
(06:24):
Sometimes it is, sometimes itisn't.
But you can't just be under thisillusion that AI is something
magic that you can just sprinkleon the problem or on your
business, and it's going to makeeverything all right.
It doesn't work that way.
I recently completed a projectwith the US Small Business
Administration, and we wereassessing the current state of
(06:47):
AI for small businesses.
And what we found is that manycompanies don't know where to
start.
So, what is the fastest way forthe leadership of these small
and medium-sized companies toagree on where they start with
AI?
SPEAKER_00 (07:05):
Yeah, well, I'll continue based on what we were
just talking about.
And I I've argued a lot that thestarting place is your
customers, is figuring out whatdo they need, what do they want,
how would they like to engagewith you.
So here's an example.
Let's just use customer supportas an example.
(07:28):
Uh, you might have some sort ofhelp desk, you might have a chat
bot that uh helps, you know,with whatever support questions
that you have.
You might find out that yourcustomers don't want to engage
with a chat bot.
They would rather call you.
Now there's an expense that isin there, but the starting point
(07:48):
here is talking about it withyour customers and figuring out
what they're okay with, whatthey're not okay with, and
starting to build strategiesaround that.
Another example of this is ifyou're familiar at all or have
seen it, or maybe you'veexperienced yourself and you've
gone to, say, like a Taco Bell,and now the drive-throughs are
(08:08):
starting to have AI assistance.
So you could order your foodthrough the drive-thru.
Now they've started to rollthose out really without a lot
of testing or talking with thecustomers to see if the customer
even likes it or not.
And in one case, what theydiscovered was that a gentleman
went through the uh drive-thruand ordered like 13,000 cups of
(08:31):
water and kind of broke thesystem.
And so Taco Bell is nowreevaluating whether or not this
is the best approach for them.
So it's an example of how, youknow, like you would do in any
other area of your business.
You would do some surveys, youwould talk to your customers,
you would find out what theirinterests are, what they're okay
(08:52):
with, what they're not okaywith, and start to build a
strategy around that.
So that's one area that youcould focus on.
And just a quick secondary waywould be to do a readiness
assessment test.
You know, look at your maturityof your organization.
Uh, do you have leadership whois saying, yes, this is
(09:15):
something that we want to do?
Do you have somebody within yourteam that is going to be uh in
charge of this?
Uh, have you identified some uh,say, pilot programs or some sort
of business objective that youwant to start small with?
So between those readiness andvalidations and maturity
assessments and talking withyour customers, uh, you should
(09:39):
be able to get a pretty goodidea as to where you can start.
SPEAKER_02 (09:44):
Yeah, I agree.
Those assessments are soimportant.
The last thing you want to do isjump into an AI implementation
and then make the customerexperience worse than it was
before, right?
The Taco Bell example being one.
You gotta be ready for it.
You have to make sure you havethe right governance processes
in place.
(10:05):
You know, testing is going to beabsolutely critical.
Make sure the system is readyfor production before you put it
out there.
There's a lot that goes into it,and I don't know that people
necessarily appreciate all that,especially, you know, we're
hearing that there's a lot ofpressure coming from the board
and from the CEO andorganizations are just trying to
(10:26):
figure out what do I do withthis, right?
How do I move forward with AI?
But I personally think it'sbetter that you do nothing than
to rush into something that'sgoing to you know be disastrous
from a customer standpoint.
So I'd like to ask Zina for heropinion.
(10:47):
Zina, what advice would you havefor companies that want to
implement AI but don't knowwhere to start?
SPEAKER_01 (10:57):
That's a great question.
I'd suggest starting with aclear understanding of the
business goals and challengesfirst, then identifying specific
use cases where AI can add valueis crucial.
It's also important to startsmall, experiment, and scale
gradually while ensuring there'sproper data governance and
talent in place.
SPEAKER_02 (11:17):
Yeah, so I think Zina is aligned with everything
that uh we talked about.
And she added starting small, Ithink that's another very good
advice, right?
Find a small opportunity withinyour business process.
Maybe it's uh a division, maybeit's a customer that you have a
very strong relationship with,and you tell them, you know,
(11:38):
this is still in testing mode,and you work with that customer
to work out the bugs.
So starting small is is really agood idea here.
Now, Chris, we hear a lot abouttransformation.
It's it's a big word, you know,everybody uses this word, but
how do you define AItransformation in practical
(12:01):
terms that leaders can act ontoday?
SPEAKER_00 (12:04):
Yeah, I think you touched on it.
Transformation.
You know, that you could removeany word that's before it,
digital transformation.
You know, I wrote a book calledCustomer Transformation, AI
transformation, remove whateverthe starting word is and just
focus on transformation.
And it's pretty much the sameacross the board.
(12:26):
And what you're really trying todo is figure out a way to adjust
your organization, you know,change management has been
always been a thing.
Uh, you know, how do you gothrough a process by which you
can change and transform yourorganization to align with
whatever the current thing is?
(12:47):
And so today it just happens tobe AI.
And as we've been talking about,like, where do you start?
And absolutely start small,experiment, find something that
uh is meaningful and valuablefor your organization.
But that transformation processwill include things like
training your staff,implementing new guardrails or
(13:11):
ground rules for how you'regoing to engage with AI.
And so even though thetechnology is changing, the
process of transformation withinan organization culturally is
really no different.
Like if we were to simplify thisand say, uh, you know, you have
(13:31):
a company and you have a brandand you've and you have a
website and you've decided, youknow what, we're going to change
our brand.
We're going to update our logoand we need to update our
website.
You're going to go through aprocess by which you do that.
And of course, part of it'sgoing to be communicating out to
the team why you're doing it andhow you're doing it and who is
(13:53):
doing it.
And AI transformation isbasically the exact same
process.
The only difference is it wrapsaround all of those policies and
procedures and reasons andrationales for doing it with AI.
SPEAKER_02 (14:08):
Yeah, so you touched on uh a number of items that are
absolutely important, right?
You talked about changemanagement, you talked about
communication, you talked abouttraining.
So this is no different than anyother big initiative that we try
to implement on organizations.
If you don't have the propergovernance processes in place,
(14:29):
if you don't have the rightproject management capabilities,
if you if the organization isnot ready from a change
management standpoint, it's notgoing to go well, right?
So let me ask Zina if she hasany specific questions for you.
So, Zina, uh given the depth ofexperience that Chris has in
(14:51):
leading AI transformation, whatquestion would you have for him?
SPEAKER_01 (14:56):
One compelling question might be
balance the need for rapid AIinnovation with the ethical
considerations and the potentialimpact on employees in society?
That could really stark aninsightful discussion.
SPEAKER_00 (15:11):
That is a very good question.
Yeah, I love that question.
So, you know, when we thinkabout speed versus control, you
know, maybe that's a simple way,you know, ethics obviously is
part of that control, governanceis part of that.
You know, speed, there'sobviously a desire to get to
market fast.
(15:31):
And uh, you know, and and weknow that when we go too fast,
we tend to encounter potentialproblems, right?
You're developing a softwaresolution and you just go to
market without testing it, andthen all of a sudden there's a
major bug that exposes all ofyour customers' data.
(15:51):
Uh, that could have beenprevented if you would have just
slowed down and did a littletesting on it, right?
And and these things do happen.
Uh, in our desire to be first,we often forget about the
critical, important smallthings.
And so uh the balance is thatyes, you can still move at a
(16:13):
very fast pace, but you can't uhavoid or ignore those core
elements that you would normallycarry out, you know, quality
assurance, bug checking, privacyof your consumer information,
all of those things are stillrequired.
And the beauty of artificialintelligence is that it can also
(16:37):
provide you with some frameworksor ground rules for ensuring
that you don't forget aboutthem.
And so as you're beginning toput together a plan for how to
execute and go to market, makesure that you are asking, you
know, core questions like let'smake sure that I don't forget
all of these other things thatare around us.
(17:00):
And I do want to touch one morething as far as just generally
the ethics are concerned.
Yeah, I do think that in ourrush to build incredible
powerful systems, we arenegating that human element.
And so we are seeing things likeforms of addiction beginning to
(17:22):
materialize due to artificialintelligence.
We are seeing that AI isinferring confidential and
private information about eachother, partly because we're
sharing that information withthe AI machines, and now it is
able to come to otherunderstandings about us, whether
(17:45):
we share that to the system ornot.
And these are really challengingproblems.
Uh, we see problems in theeducational system with students
using AI to cheat.
We are seeing uh flattenings ofour language, where common
consistency in our anddifferentiations in our voices
(18:06):
are starting to decline becausecritical thinking skills are
going away.
All of these core human traitsare being challenged.
And and I wouldn't argue thatthis is purely to rush to
market.
I would argue that this is moreabout pure money, right?
The the desire and greed formore uh market and capital
(18:32):
domination is indirectly causingthese, you know, these serious
ethical concerns.
And that is something that wereally need to address.
SPEAKER_02 (18:43):
That's a really good point.
There are so many ethicalconsiderations that we need to
account for as we continue todevelop and roll out AI.
You know, there's the issue ofprivacy, the issue of uh biases
and accountability andtransparency and explainability,
(19:03):
and the list goes on and on.
There are a lot of greatconcerns, and there are also a
lot of great people thinkingabout these issues and
advocating for AI that isresponsible, that follows the
democratic processes, thatfollows the rule of law, that
takes into account humandignity.
(19:23):
So I've been doing some researchwork on AI policy, and I'm
seeing a lot of great peopledoing tremendous work trying to
advocate for ethical andresponsible AI, and that's
that's really good to see.
Now, Chris, I'm sure you'veheard about this MIT study that
(19:44):
shows that 95% of AIimplementations fail.
Now, from your viewpoint, whatcan companies do to increase the
probability of success when uhimplementing AI?
SPEAKER_00 (20:01):
Well, there's two core challenges in this.
You know, there's been a lot ofconversation about the
legitimacy of that particularMIT report.
I'm gonna put that aside.
We'll just call it what it is.
There is some percentage ofbusinesses that are failing in
their AI implementation, right?
(20:22):
I don't care if it's 5% or 95%,or if it's 200, you know,
businesses or 5,000 businesses.
There is a percentage that isfailing.
And we have to ask ourselves,why are these uh businesses
failing?
My argument is that the numberone reason why these businesses
are failing is because there isa misalignment and
(20:44):
misunderstanding of what AI is,what the capabilities of AI are,
and how they're being sold ormarketed to those capabilities.
And so when there's thisdisconnect, when you know
technology is overhyped andoverpromised and does not align
(21:05):
with actual capabilities, thenbusinesses begin to go down a
path of implementing this as asolution that's going to solve
all of our problems or fill inthe blank.
And when that doesn't happen,then there's obviously a failure
ratio of some sort.
(21:26):
Again, this is mainly becausethe understood capabilities do
not align with actualcapabilities.
In my uh book that I launched atthe beginning of this year, in
failable, I did research, andwhat I found was that the
understanding or the belief incurrent AI capabilities versus
(21:51):
actual capabilities is about aseven-year gap.
And the reason for thisseven-year gap is primarily due
to media and movies, TV.
And I I've said that you know,when we learn about the
capabilities of AI by watchingthe Terminator, then we are not
(22:14):
living in a reality of actualcapabilities.
And then you tap on top of that,that you have marketing out
there shoving down everybody'sthroat that AI will solve these
problems, or AI is capable ofdoing these things when it
can't, then yes, there's goingto be a lot of disappointed
(22:36):
business leaders out theretrying to figure it out.
And so that's one reason, whichI argue is probably the biggest
reason.
But yeah, there's there's otherminor things, you know, just in
terms of, again, not aligning acore business value principle,
you know, customer value uhopportunity.
(22:59):
Uh, you know, when you're notlooking at it from a business
lens and you're just trying tohop onto the bandwagon, that's
also going to cause problems.
SPEAKER_02 (23:07):
You made such a good point.
Misunderstanding of AI issomething that I think is one of
the causes for these failures.
People just don't understand thecapabilities of the technology.
And I think there's a lot ofconfusion out there.
People are confused, they don'tunderstand it, they don't know
(23:28):
where to start, they hearthings, like you said, you know,
marketing is pushing all thisstuff, but they don't really
understand exactly how thesetools can help them.
I think there's a lot of uhpeople out there that are
looking for guidance and lookingfor people like you that can
help them steer them in theright direction, right?
I think is what they're lookingfor.
(23:49):
Now, to add to this confusion,just to very briefly touch on
this subject, I wanted to talkabout Agenc AI.
And I know that you have donesome great work with Agenc AI.
So can you tell our audiencejust in 20 seconds what is
agentic AI and is it real?
Is it working today?
SPEAKER_00 (24:11):
Uh, it is real, it is not what people think it is.
And agentic AI basically, in thesimplest terms, is a toolbox of
actions or a toolbox of toolsthat systems can use to perform
various actions.
You know, so uh think about itlike you might have an agent, a
(24:33):
system that wants to perform atask.
And we'll say something likebook a flight.
You know, you want to go on atrip.
And so you ask the agent, canyou help me book a flight?
Well, the agent goes through theprocess of thinking about it.
The actual booking, you know, Isay book a flight, the book or
(24:54):
the reservation, that's agentic,and that's all it is.
So think about it like atoolbox.
It's a a group of things that itcan do, action-based adverbs if
you want.
And then from there, nothingelse around that is agentic.
SPEAKER_02 (25:12):
I think we're going to see tremendous progress in
this particular area.
I do have concerns about ourability to test these agentic AI
systems because they're notdeterministic.
So, you know, our mindset comingfrom the way we have tested
software in the past is notgoing to work in the world of
(25:32):
Agentic AI.
So it'll be interesting to seehow that evolves.
So, Chris, thank you so much forbeing a guest on the show today.
You have uh a wealth ofknowledge, and it's just uh
wonderful to be able to havethis conversation with you.
Thank you for being here withus.
SPEAKER_00 (25:51):
Absolutely.
My pleasure.
SPEAKER_02 (25:53):
And Zina, thank you so much again for being a great
co host.
SPEAKER_01 (25:58):
Of course, David.
It's always a pleasure to bepart of the conversation.
I'm looking forward to anotherfantastic episode.
Let's make it great.
SPEAKER_02 (26:06):
Sounds good.
Thank you, Zina.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com