May 2, 2025 • 21 mins
What if the secret to harnessing AI in your business lies in strong governance and collaboration?
In this episode, host Andreas Welsch speaks with Maxim Ioffe, Director Global IA CoE at WESCO, about the critical role of governance in AI adoption. They explore how companies can strategically align technology with business goals while fostering an innovative culture.
Maxim shares invaluable insights on developing a comprehensive AI mission, ensuring that governance frameworks are in place, and emphasizing the importance of employee education in AI integration. As organizations face challenges around ethics, security, and the risk of rogue AI behavior, this episode offers a roadmap for transforming AI from a buzzword into a powerful business ally:
- How can IT organizations become an innovation engine rather than be the breaking pads?
- What strategies are effective for raising awareness of data privacy and security with AI?
- What does this new reality of consumer AI tools mean for IT leaders?
- Who should own the AI governance, and who needs to be around the table?
Whether you're a tech leader or a business executive, the actionable strategies discussed will help you create a thriving environment for AI adoption. Are you ready to dive into the world of AI governance and innovation?
Don’t miss this episode—tune in now for groundbreaking insights that could shape your business's future!
Questions or suggestions? Send me a Text Message.
***********
Disclaimer: Views are the participants’ own and do not represent those of any participant’s past, present, or future employers. Participation in this event is independent of any potential business relationship (past, present, or future) between the participants or between their employers.
Level up your AI Leadership game with the AI Leadership Handbook:
https://www.aileadershiphandbook.com
More details:
https://www.intelligence-briefing.com
All episodes:
https://www.intelligence-briefing.com/podcast
Get a weekly thought-provoking post in your inbox:
https://www.intelligence-briefing.com/newsletter
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Andreas Welsch (00:00):
Welcome to What's the BUZZ?, where leaders
(00:02):
and hands-on experts share howthey have turned hype into
outcome.
Today, we'll talk aboutestablishing a strong governance
between technology and businessteams, and who better to talk
about it than someone who'sactively doing that in their
business.
Maxim Ioffe.
Hey, Maxim, thank you so muchfor joining.
Maxim Ioffe (00:18):
Thanks for having me.
Andreas Welsch (00:20):
Wonderful.
Why don't you tell our audiencea little bit about yourself, who
you are, and what you do?
Maxim Ioffe (00:25):
Sure.
My name is Max and I driveGlobal Intelligence, center of
Excellence for is a Fortune 200wholesale distribution company.
It's about 20 billion dollarsannual.
We are partnered with 50,000suppliers, hundred thousands of
customers, and that's millionsof products.
With that scale, there is noshortage, automation, and my
(00:46):
mission is to help everybody andanybody to be more effective and
more efficient.
Andreas Welsch (00:51):
That's awesome.
I love that.
And you've been on the showbefore and we talked about a
pretty similar topic, but notquite the one about governance.
So we wanted to make good on, onthat now as well.
Now, Maxim, what do you say ingood old fashion, should we play
a little game to kick thingsoff?
Maxim Ioffe (01:06):
Sure.
Andreas Welsch (01:07):
All right, so let's do this.
If AI were a, let's see.
If it were a plant, what wouldit be?
60 seconds on the clock.
Go.
Maxim Ioffe (01:22):
All right.
It's probably going to besomething from a cacti family.
It's going be difficult toclimb.
It's gonna be d difficult tohandle, but if you handle it
just right, the Greek pears arepretty delicious.
Andreas Welsch (01:37):
Wonderful.
I love that.
Really good answer.
And there they're all theselittle spikes if you don't know,
that they exist, it, it can getpretty interesting.
Speaking of interesting and someof the spikes that I think
business and technologists.
Feel on a regular basis, how dowe even work together, right?
We want to do AI, we need to doAI, but how do we actually get
(01:58):
everybody around the same tableand move in the same direction?
I keep hearing these concernsfrom IT leaders when it comes to
adoption.
We need to look at the data.
We need to mitigate risk, weneed to keep an eye on our costs
and how can I become aninnovation engine rather than
being the braking pads if youwill.
(02:20):
I'm curious, what are you seeingin your role?
How can IT/ AI automationleaders become an innovation
engine while still keeping allthese things in view.
Maxim Ioffe (02:32):
That's a great question.
And I don't think there is onesize fits all answer that would
cover every use case, everypossibility, everybody and
anybody.
That's just not realistic.
But I think fundamentally wehave a couple of pillars there
that needs to apply to buildinga successful model.
Just one is definitely strategicand the strategy starts with
(02:53):
something as basically as amission.
What is the mission of AI?
And broadly speaking, theanimation program that we're
trying to run.
And I'll do a simple example.
And a lot of missions seem to begoing about, let's build a
autonomous enterprise.
Let's build some sort of a faildriving engine.
And then you turn around at thecompany and that corporate
(03:14):
values are talking about peopleas the greatest asset.
And how do you think ourgreatest fail?
If we say our mission is tobuild an autonomous enterprise,
right?
So crafting that mission,crafting of golf that resonates
with business, that resonateswith employees, that resonates
with everybody and makes senseis a big part of the strategy.
The other part of the strategyis defining what are we after,
(03:36):
what kind of, how do wecalculate value?
I don't wanna use the word ROI.
ROI seems to be a little bitlimiting to some of those
things, but at the end of theday, it is a financial
calculation.
How do we bring the financingand tell us, Hey, help us to
define the appropriate levers,and how do you cal calculate us?
So you have the strategy aroundit.
(03:58):
And then the other part of thestrategy is probably the,
process selection and eventevangelizing.
So you keep the stream of theprocesss, right?
It's probably not that hard tofind the first pro project to
work on.
Maybe first 2, 3, 4, how to getinto hundreds.
In our program we had a smallmilestone.
We have one in 500 ideas in the.
(04:19):
Automation hub that are eitherimplemented or being worked on.
And 500 ideas is not a end tothe game.
And just the beginning, we'regonna keep growing, we're going
to keep more, adding more andmore.
And frankly, that's 500 iswhat's live now.
There is a number that werealready automated and archived
because they outlive their usuallifespan and there is nothing
(04:40):
wrong about that as well.
So that's your strategy piece,right?
The second pillar is probably,funding.
How do you go around funding?
Because frankly, strategywithout funding is more about
dreams than reality.
You can dream up whateverstrategy, but if you don't have
the budget to build it, it's notgonna build.
And then on top of that, they'reinterconnected.
(05:00):
You have the budget and a lot oftime the budget is this.
Yeah, you might have a strategythat requires many millions of
dollars, but the budget is muchsmaller.
So be it.
You got to adjust the strategyto meet the budget.
And I think the third pillar isgovernance.
How do you build a frameworkwhere every idea becomes handle
as a factory?
It's not a one off.
(05:21):
I'm trying to approach this ideathis way, that idea, that way.
Back to our plant analogy.
There are a lot of spikes and alot of turns that if you are not
building that process where youdrop each one of them in a
consistent manner and you knowexactly how to do it so you
don't run into those objects,life is gonna be a lot easier.
And to me, those retailers thatmake a difference between a
(05:44):
program and a science exerciseor a one up attempt.
Andreas Welsch (05:49):
That's right.
Great point.
Especially around making it aprogram and, not just a science
experiment.
And, I feel in many cases we'relikely going to see more of
these science experiments ifpeople are not yet following
that, that three step approachbecause.
Hey, here's a new technology.
We need to try that out.
We've seen that with Gen AI.
We're now seeing that withagentic AI, although I think
(06:12):
that's still in it in itsinfancy when it comes to
evaluation.
Your words of wisdom definitelyare important here now.
I also think it's harder thanever to put a proper governance
around using AI in, in abusiness.
You have team members that mightalready use AI apps on their
phones.
They might be puttingconfidential data in there and
what have you.
(06:32):
Especially if you try to clampit down from a corporate point
of view.
And you say you're not allowedto use gen AI because we're
afraid you put data there.
Chances are people are startingto use apps on their own phones
and they're still going to putcompany data in there.
And, training can be aneffective aspect to, to raise
that awareness.
But how do you see IT leadershandling, that part of AI
(06:54):
adoption or reckoning that it'shere?
How do we deal with that now?
Maxim Ioffe (06:58):
That's a great question.
And again, I if one size fitsall, but I would start with
something basic as employee.
I just hang up the phone, right?
Prior to this when I was talkingto new employees and I was
talking about the drop of thefuture, I was talking about,
Hey, here's what we can do.
Here's how automation can helpyou.
(07:19):
Hopefully those bits ofeducation coming from me, coming
from my peers, I'm definitelynot the only one in the company
doing that.
Should help the employees tounderstand where it is coming
from, where business is comingfrom, where everybody's coming
from, and the lines of caution.
Yes, there are tremendousbenefit to AI.
Yes, there is significant riskto AI and there is security,
(07:41):
there is compliance, there isbiases.
There is that availability.
But as you educate theemployees, couple of things
happen.
One is the pipeline of theopportunities become more
realistic.
How often do we get approachedby an employee saying, I.
I want to build AI model thatwill do something.
And you start asking thequestions, do we have the data?
No, not really.
(08:02):
Do we have the process?
No, we don't really have aprocess.
Do we have good understanding ofhow things connect?
Your AI guy, you should be ableto figure it out.
Oh, it doesn't work that way.
On the flip side of it to yourpoint, cannot be.
So they need to be the, that, inmy opinion, at least, they need
to be a very consistentframework.
Here are the things that it islooking for before they allow an
(08:25):
AI solution.
And it has to do with security.
It has to do with privacy, ithas to do with handling data,
has to do with every aspect thatIT and security and
infrastructure are concernedabout.
It also has to do with toolproliferation.
How many IDP software programscan we have everybody?
(08:46):
Seems to have an IDP solution.
If we have 30 companies doingintelligent document processing
for us, are we gaining anythingor do we wanna standardize the
mechanical?
So all those different thingsare important.
All those different things comefrom it.
The financial calculation needsto be, again, the same yard
stake.
This is back to my governancestandpoint.
If you have that yard stake thatsays this program, if we do this
(09:09):
project, we're going to save X.
That kind of determines thebudget for building the
implementation.
The implementation should not bemore than that.
Otherwise it's a negative ROIimplementation right there.
And then if you get all thosepieces together, you have the IT
government saying, here's whatwe need in order to allow it.
You have business employees whoare educated enough to bring the
(09:30):
real examples and to be able toseparate the high from reality,
and you have the consistentmetrics on the value that it
brings problem becomes a littlebit easier.
Again, it's the differencebetween the science, exercise
and yelling and the loudestvoice gets the project versus we
approach it methodically.
(09:51):
We know the ROI.
We know the governance, we knowthe strategy, let's go build it.
And if it all fits together, itbecomes that factory setting.
Andreas Welsch (10:00):
So how do you do that?
How do you get into that rhythm?
Do you set up a formal program?
Do you meet on a regular basis?
Who do you invite?
What do you review?
Maxim Ioffe (10:08):
I don't know if our example would work for every
company.
What we do is we do meet onregular basis.
We do.
Yeah, prior to intakes workedvery diligently and very hard to
build a role of engagement, andI will speak from my experience
on intelligent animationprogram.
What on the RPA side, since I'vebeen doing that since 2019, we
(10:30):
met with Build the Framework.
We tweak the framework, we gotthe framework.
Every new idea I can get itthrough approvals in a matter of
minutes.
Literally, we're not there withAI yet.
It takes longer, but that's theaspiration.
There got to be that matrix thatyou need to populate to say,
does it fit?
And if the answer is yes, itfits on every account, it just
(10:52):
goes.
If the answer is yes, except forthis one point.
We need to discuss just this onepoint, but you got to make sure
that the boundaries are set.
The boundaries are don't move,and the boundaries are the
boundaries.
Business needs to respect theboundaries.
I'm not IT security person.
I may not understand all theimplications.
(11:12):
I cannot speak to everythingaround security.
So I trust the professionals todo that.
They tell me what needs tohappen.
I make it happen in order tomove forward.
That's part of the governanceframework.
Andreas Welsch (11:25):
And how high do you bring that transparency in
the organization if youcollaborate with your business
stakeholders to also show anddemonstrate the value that, that
your teams are bringing from anautomation, from an AI point of
view.
Maxim Ioffe (11:37):
I think that there are two answers here.
One is top down.
I do use every opportunity.
I direct reports to talk programour existance and what we can
deliver.
I do the same thing, bottom upapproach, where most of the
ideas come from the employees.
And I want the employees to beour investors.
(11:58):
I want the employees to go totheir managers, to their
executives and say, look at thecool automations we built.
Look at this cool agent orwhatever solution it is that we
implemented.
Look how it helps us to hit thegoals.
I will not be able to speak thelanguage of every executive that
just not in the car.
I'm not an HR professional.
(12:18):
I'm not a legal professional.
I'm not a sales professional.
I'm not a supply chainprofessional.
I know enough to be.
Somewhat tender, but I'm not aprofessional there.
So let the employees who do thework give my spoke people, they
will use the right language,they will communicate the ROI
much better.
My work is to provide thegovernance, the framework, the
training, the implementations,and the support.
(12:41):
But let the others do the sale.
I think that's a lot moreefficient and effective.
But again, I might be in aunique position when I'm able to
do it that way.
Andreas Welsch (12:49):
Yeah I really love it, especially the part
about empowering employees toidentify where the opportunities
in our business process, in thearea that I'm deeply familiar
with, and I remember last timeyou were on, on the show, we
talked about from a center ofexcellence point of view, what
do you need to do?
So your pipeline of ideas, thatdoesn't dry up, that kind of
goes hand in hand with thattopic.
(13:10):
Absolutely.
You mentioned governance and whoshould own that AI governance
and what other teams do you needto have around the table?
I think you already mentioned afew, but are there others in
finance or HR that you have inyour business program?
Maxim Ioffe (13:24):
Right there, there are many players out there and I
don't have a good answer.
Who should own it?
I think it depends on thecompany.
Some companies it's better withIT, some companies it might sit
better with some sort of a valuerealization office and CFO
organization.
It might be the executivebranch.
Doesn't really matter where itsits.
It's all about thatcollaboration.
'cause ultimately there is a lotof input from the it's the
(13:46):
infrastructure, it's thesecurity, it's the data.
But there is also input from thedata governance and data
modeling and people that buildAI models and.
People who understand the dataand the technology and there is
input from the business.
And ultimately the ideas have tostart from the business.
Business need to want toautomate.
(14:06):
Coming to the business saying,Hey, from ITM here to automate
your work is probably not ideal.
Business is going veryapprehensive.
A lot of fears of job protectionare out there and a lot of them
are probably unfounded, but itdoesn't matter.
They're still figures and westill need to deal with change
management.
And I choose to deal with changemanagement from the bottom up.
(14:27):
So if the business brings theideas to us and we want to do
it, not the business leader, butthe employees within the
business line, I know that theydon't need to change management
employees already happy to doit.
Maybe we can steer it, maybe wecan help them to scale it.
Maybe we can help them to thinkbigger.
But to me that's how do yourecord the value in a consistent
(14:48):
manner so you don't have an AIproject that saves millions of
dollars, and those millionsnever hit the p and l.
They never hit any financialstatements.
So those are.
Imaginary dollars.
It's fine to report it that wayif that's what the fellowship
organization wants us to do.
But most of the time they wantus to report on the real dogs
and let's report within thenumbers and the guidelines that
(15:09):
they want us to report.
Within.
Saves time save frictions makeslife easier.
And employee education.
It might be sitting within abusiness within it or somewhere
else, but to me it's one of thestaples.
If you don't have employees knowwhat to ask for, they will not
ask for the right things.
If you don't have the employeeswho know how the technology
works, they will come withunrealistic expectations.
(15:32):
Asking for the wrong thing,having realistic expectations is
never useful.
Andreas Welsch (15:37):
Wow, there was so much good information in
there.
Let me let me pick out a fewthings.
What I heard you say was, weneed to report on hard actual
savings, not just the funnymoney that we shift around
between business units but howare we actually making impact to
the business, to the P&L that'swhat our business wants and we
(15:58):
can report it and we should.
So we showed the real impact.
I love that because I think toomany times we still see the
sandbox projects or pilots.
Let's figure out what we do withAI and then maybe we roll it out
and it's actually not that easy.
But to your point, if you startwith that from the very
beginning, align it with yourbusiness strategy.
Take something that'smeasurable, that solves that
problem, right?
(16:18):
You don't even run into it.
The other part I wanted to askis, how do you get employees to
that point that they are awareof?
What can the technology actuallydo?
What is available at ourdisposal?
And could this even be somethingthat we can solve with
technology?
How do you do that?
Maxim Ioffe (16:34):
It's all about that continuous education.
I'm not a marketing person, butI've heard something talking
about pre attaches to theperson.
You need to pass thatinformation at least three times
before it ings.
I took that to heart and I triedto do it more than three times.
Every opportunity I get to talkto employees to present, it
could be a lunch and learn, itcould be five minutes.
(16:56):
During the town hall, we haveautomation minute, if you will,
where I just go in and present.
I offer it every call,everything I do.
Hey, if you want to do targetedpresentation for your employees,
happy to do it.
It's up to us who understand thetechnology and can speak speak
about it not in the IT language,but in the language that
(17:16):
business can relate to andunderstand.
Be the ambassadors of thetechnology.
It'll not happen on its own, andemployees have bombarded by
advertisement and thatadvertisement with all the right
intentions.
Create a very unrealisticpicture.
We are going to put AI in placefor your organization that will
reconcile every invoice.
(17:36):
We're going to put AI into theyour organization that is going
to take care of all therecruitment.
P two P is going to be handledby our AI agent.
How often do we see that onLinkedIn feeds?
How often do we see that duringthe presentations, equipping the
employees about it?
Talking about it?
And I, again, just right beforethis call, I was talking about
Gartner's hype curve to newemployees.
(18:00):
And we ask, say, look, when youget a new presentation about
technology.
Think where it falls into thathype.
Is it on peak of the inflatedexpectation or is it within the
trout of disillusionment?
It doesn't have to be thatterminology, but it gives you
some sort of a framework to talkabout, look, this is what is
needed.
This is how you identify thereality versus hype.
(18:22):
By the way, employee, if youbring to me something that
sounds great, but we can neverimplement it.
Nobody wins, right?
So if you know that it all logs,let's talk about it upfront and
see how we can deal with it.
There's no magic in thattechnology.
Again, a lot of folks try topretend that AI is some sort of
a magical tool that does a lotof things.
(18:43):
In a way it is.
But the reality.
Teaching the employees to checkfor that foundation is very
effective from a standpointthat, again, we're not experts.
Typically, the people whoimplement AI are not experts in
the business problem.
Yeah.
So we cannot say if thatfoundation is solid, we need
employees to help us with it.
Andreas Welsch (19:00):
That's awesome.
Maxim, there's been so manygreat pieces of information, so
many nuggets that you've sharedin the last 20, 25 minutes.
And I was wondering if you cansummarize the key three
takeaways for our audiencetoday.
Maxim Ioffe (19:13):
I wish I had your LinkedIn post from yesterday
open.
You summarized it really well.
Where the technology is goingand building 10 bots is not a
good strategy.
Three enough, 10 is not a goodstrategy.
The strategy needs to be aboutthe technology, measurable
impact of that technologies andthe strategy needs to be
something that you are willingand going to execute and include
(19:33):
at North Star.
Where do you see it going andhow do we get there from where
we are to where we wanna bewithin that strategy?
We wanna make sure that we aresupported by funding and we know
how to be funded, how to executeit.
And then the governance.
The governance is a key from astandpoint that you cannot treat
each project as a one up.
It has to be an autopilot.
(19:54):
And if it's an autopilot, thereis no friction between it and
business and everybody else.
Everybody expects and respectsthose boundaries and follows
them.
Why become easy?
You spend less time arguing,more time doing, and that's the
goal.
And the last thing I will say,maybe it's number four, is
measure twice.
Cut once.
Prep, the prep, all the things.
(20:15):
Figure out the strategy, thegovernance, everything else
before you buy the licenses,before you sign the agreement,
before you start implementing.
It makes life a lot easier and alot less mistakes are again,
getting made.
So thus would be my summary.
Andreas Welsch (20:30):
Fantastic.
Thank you so much folks.
We're at the end of the show.
Thank you so much, Maxim, forjoining us today and for sharing
your experience with us.
It was a pleasure having you onagain.
Maxim Ioffe (20:40):
My pleasure.
Thank you so much for having me.
Andreas Welsch (20:42):
Wonderful.
Welcome to What's the BUZZ?, where leaders
(00:02):
and hands-on experts share howthey have turned hype into
outcome.
Today, we'll talk aboutestablishing a strong governance
between technology and businessteams, and who better to talk
about it than someone who'sactively doing that in their
business.
Maxim Ioffe.
Hey, Maxim, thank you so muchfor joining.
Maxim Ioffe (00:18):
Thanks for having me.
Andreas Welsch (00:20):
Wonderful.
Why don't you tell our audiencea little bit about yourself, who
you are, and what you do?
Maxim Ioffe (00:25):
Sure.
My name is Max and I driveGlobal Intelligence, center of
Excellence for is a Fortune 200wholesale distribution company.
It's about 20 billion dollarsannual.
We are partnered with 50,000suppliers, hundred thousands of
customers, and that's millionsof products.
With that scale, there is noshortage, automation, and my
(00:46):
mission is to help everybody andanybody to be more effective and
more efficient.
Andreas Welsch (00:51):
That's awesome.
I love that.
And you've been on the showbefore and we talked about a
pretty similar topic, but notquite the one about governance.
So we wanted to make good on, onthat now as well.
Now, Maxim, what do you say ingood old fashion, should we play
a little game to kick thingsoff?
Maxim Ioffe (01:06):
Sure.
Andreas Welsch (01:07):
All right, so let's do this.
If AI were a, let's see.
If it were a plant, what wouldit be?
60 seconds on the clock.
Go.
Maxim Ioffe (01:22):
All right.
It's probably going to besomething from a cacti family.
It's going be difficult toclimb.
It's gonna be d difficult tohandle, but if you handle it
just right, the Greek pears arepretty delicious.
Andreas Welsch (01:37):
Wonderful.
I love that.
Really good answer.
And there they're all theselittle spikes if you don't know,
that they exist, it, it can getpretty interesting.
Speaking of interesting and someof the spikes that I think
business and technologists.
Feel on a regular basis, how dowe even work together, right?
We want to do AI, we need to doAI, but how do we actually get
(01:58):
everybody around the same tableand move in the same direction?
I keep hearing these concernsfrom IT leaders when it comes to
adoption.
We need to look at the data.
We need to mitigate risk, weneed to keep an eye on our costs
and how can I become aninnovation engine rather than
being the braking pads if youwill.
(02:20):
I'm curious, what are you seeingin your role?
How can IT/ AI automationleaders become an innovation
engine while still keeping allthese things in view.
Maxim Ioffe (02:32):
That's a great question.
And I don't think there is onesize fits all answer that would
cover every use case, everypossibility, everybody and
anybody.
That's just not realistic.
But I think fundamentally wehave a couple of pillars there
that needs to apply to buildinga successful model.
Just one is definitely strategicand the strategy starts with
(02:53):
something as basically as amission.
What is the mission of AI?
And broadly speaking, theanimation program that we're
trying to run.
And I'll do a simple example.
And a lot of missions seem to begoing about, let's build a
autonomous enterprise.
Let's build some sort of a faildriving engine.
And then you turn around at thecompany and that corporate
(03:14):
values are talking about peopleas the greatest asset.
And how do you think ourgreatest fail?
If we say our mission is tobuild an autonomous enterprise,
right?
So crafting that mission,crafting of golf that resonates
with business, that resonateswith employees, that resonates
with everybody and makes senseis a big part of the strategy.
The other part of the strategyis defining what are we after,
(03:36):
what kind of, how do wecalculate value?
I don't wanna use the word ROI.
ROI seems to be a little bitlimiting to some of those
things, but at the end of theday, it is a financial
calculation.
How do we bring the financingand tell us, Hey, help us to
define the appropriate levers,and how do you cal calculate us?
So you have the strategy aroundit.
(03:58):
And then the other part of thestrategy is probably the,
process selection and eventevangelizing.
So you keep the stream of theprocesss, right?
It's probably not that hard tofind the first pro project to
work on.
Maybe first 2, 3, 4, how to getinto hundreds.
In our program we had a smallmilestone.
We have one in 500 ideas in the.
(04:19):
Automation hub that are eitherimplemented or being worked on.
And 500 ideas is not a end tothe game.
And just the beginning, we'regonna keep growing, we're going
to keep more, adding more andmore.
And frankly, that's 500 iswhat's live now.
There is a number that werealready automated and archived
because they outlive their usuallifespan and there is nothing
(04:40):
wrong about that as well.
So that's your strategy piece,right?
The second pillar is probably,funding.
How do you go around funding?
Because frankly, strategywithout funding is more about
dreams than reality.
You can dream up whateverstrategy, but if you don't have
the budget to build it, it's notgonna build.
And then on top of that, they'reinterconnected.
(05:00):
You have the budget and a lot oftime the budget is this.
Yeah, you might have a strategythat requires many millions of
dollars, but the budget is muchsmaller.
So be it.
You got to adjust the strategyto meet the budget.
And I think the third pillar isgovernance.
How do you build a frameworkwhere every idea becomes handle
as a factory?
It's not a one off.
(05:21):
I'm trying to approach this ideathis way, that idea, that way.
Back to our plant analogy.
There are a lot of spikes and alot of turns that if you are not
building that process where youdrop each one of them in a
consistent manner and you knowexactly how to do it so you
don't run into those objects,life is gonna be a lot easier.
And to me, those retailers thatmake a difference between a
(05:44):
program and a science exerciseor a one up attempt.
Andreas Welsch (05:49):
That's right.
Great point.
Especially around making it aprogram and, not just a science
experiment.
And, I feel in many cases we'relikely going to see more of
these science experiments ifpeople are not yet following
that, that three step approachbecause.
Hey, here's a new technology.
We need to try that out.
We've seen that with Gen AI.
We're now seeing that withagentic AI, although I think
(06:12):
that's still in it in itsinfancy when it comes to
evaluation.
Your words of wisdom definitelyare important here now.
I also think it's harder thanever to put a proper governance
around using AI in, in abusiness.
You have team members that mightalready use AI apps on their
phones.
They might be puttingconfidential data in there and
what have you.
(06:32):
Especially if you try to clampit down from a corporate point
of view.
And you say you're not allowedto use gen AI because we're
afraid you put data there.
Chances are people are startingto use apps on their own phones
and they're still going to putcompany data in there.
And, training can be aneffective aspect to, to raise
that awareness.
But how do you see IT leadershandling, that part of AI
(06:54):
adoption or reckoning that it'shere?
How do we deal with that now?
Maxim Ioffe (06:58):
That's a great question.
And again, I if one size fitsall, but I would start with
something basic as employee.
I just hang up the phone, right?
Prior to this when I was talkingto new employees and I was
talking about the drop of thefuture, I was talking about,
Hey, here's what we can do.
Here's how automation can helpyou.
(07:19):
Hopefully those bits ofeducation coming from me, coming
from my peers, I'm definitelynot the only one in the company
doing that.
Should help the employees tounderstand where it is coming
from, where business is comingfrom, where everybody's coming
from, and the lines of caution.
Yes, there are tremendousbenefit to AI.
Yes, there is significant riskto AI and there is security,
(07:41):
there is compliance, there isbiases.
There is that availability.
But as you educate theemployees, couple of things
happen.
One is the pipeline of theopportunities become more
realistic.
How often do we get approachedby an employee saying, I.
I want to build AI model thatwill do something.
And you start asking thequestions, do we have the data?
No, not really.
(08:02):
Do we have the process?
No, we don't really have aprocess.
Do we have good understanding ofhow things connect?
Your AI guy, you should be ableto figure it out.
Oh, it doesn't work that way.
On the flip side of it to yourpoint, cannot be.
So they need to be the, that, inmy opinion, at least, they need
to be a very consistentframework.
Here are the things that it islooking for before they allow an
(08:25):
AI solution.
And it has to do with security.
It has to do with privacy, ithas to do with handling data,
has to do with every aspect thatIT and security and
infrastructure are concernedabout.
It also has to do with toolproliferation.
How many IDP software programscan we have everybody?
(08:46):
Seems to have an IDP solution.
If we have 30 companies doingintelligent document processing
for us, are we gaining anythingor do we wanna standardize the
mechanical?
So all those different thingsare important.
All those different things comefrom it.
The financial calculation needsto be, again, the same yard
stake.
This is back to my governancestandpoint.
If you have that yard stake thatsays this program, if we do this
(09:09):
project, we're going to save X.
That kind of determines thebudget for building the
implementation.
The implementation should not bemore than that.
Otherwise it's a negative ROIimplementation right there.
And then if you get all thosepieces together, you have the IT
government saying, here's whatwe need in order to allow it.
You have business employees whoare educated enough to bring the
(09:30):
real examples and to be able toseparate the high from reality,
and you have the consistentmetrics on the value that it
brings problem becomes a littlebit easier.
Again, it's the differencebetween the science, exercise
and yelling and the loudestvoice gets the project versus we
approach it methodically.
(09:51):
We know the ROI.
We know the governance, we knowthe strategy, let's go build it.
And if it all fits together, itbecomes that factory setting.
Andreas Welsch (10:00):
So how do you do that?
How do you get into that rhythm?
Do you set up a formal program?
Do you meet on a regular basis?
Who do you invite?
What do you review?
Maxim Ioffe (10:08):
I don't know if our example would work for every
company.
What we do is we do meet onregular basis.
We do.
Yeah, prior to intakes workedvery diligently and very hard to
build a role of engagement, andI will speak from my experience
on intelligent animationprogram.
What on the RPA side, since I'vebeen doing that since 2019, we
(10:30):
met with Build the Framework.
We tweak the framework, we gotthe framework.
Every new idea I can get itthrough approvals in a matter of
minutes.
Literally, we're not there withAI yet.
It takes longer, but that's theaspiration.
There got to be that matrix thatyou need to populate to say,
does it fit?
And if the answer is yes, itfits on every account, it just
(10:52):
goes.
If the answer is yes, except forthis one point.
We need to discuss just this onepoint, but you got to make sure
that the boundaries are set.
The boundaries are don't move,and the boundaries are the
boundaries.
Business needs to respect theboundaries.
I'm not IT security person.
I may not understand all theimplications.
(11:12):
I cannot speak to everythingaround security.
So I trust the professionals todo that.
They tell me what needs tohappen.
I make it happen in order tomove forward.
That's part of the governanceframework.
Andreas Welsch (11:25):
And how high do you bring that transparency in
the organization if youcollaborate with your business
stakeholders to also show anddemonstrate the value that, that
your teams are bringing from anautomation, from an AI point of
view.
Maxim Ioffe (11:37):
I think that there are two answers here.
One is top down.
I do use every opportunity.
I direct reports to talk programour existance and what we can
deliver.
I do the same thing, bottom upapproach, where most of the
ideas come from the employees.
And I want the employees to beour investors.
(11:58):
I want the employees to go totheir managers, to their
executives and say, look at thecool automations we built.
Look at this cool agent orwhatever solution it is that we
implemented.
Look how it helps us to hit thegoals.
I will not be able to speak thelanguage of every executive that
just not in the car.
I'm not an HR professional.
(12:18):
I'm not a legal professional.
I'm not a sales professional.
I'm not a supply chainprofessional.
I know enough to be.
Somewhat tender, but I'm not aprofessional there.
So let the employees who do thework give my spoke people, they
will use the right language,they will communicate the ROI
much better.
My work is to provide thegovernance, the framework, the
training, the implementations,and the support.
(12:41):
But let the others do the sale.
I think that's a lot moreefficient and effective.
But again, I might be in aunique position when I'm able to
do it that way.
Andreas Welsch (12:49):
Yeah I really love it, especially the part
about empowering employees toidentify where the opportunities
in our business process, in thearea that I'm deeply familiar
with, and I remember last timeyou were on, on the show, we
talked about from a center ofexcellence point of view, what
do you need to do?
So your pipeline of ideas, thatdoesn't dry up, that kind of
goes hand in hand with thattopic.
(13:10):
Absolutely.
You mentioned governance and whoshould own that AI governance
and what other teams do you needto have around the table?
I think you already mentioned afew, but are there others in
finance or HR that you have inyour business program?
Maxim Ioffe (13:24):
Right there, there are many players out there and I
don't have a good answer.
Who should own it?
I think it depends on thecompany.
Some companies it's better withIT, some companies it might sit
better with some sort of a valuerealization office and CFO
organization.
It might be the executivebranch.
Doesn't really matter where itsits.
It's all about thatcollaboration.
'cause ultimately there is a lotof input from the it's the
(13:46):
infrastructure, it's thesecurity, it's the data.
But there is also input from thedata governance and data
modeling and people that buildAI models and.
People who understand the dataand the technology and there is
input from the business.
And ultimately the ideas have tostart from the business.
Business need to want toautomate.
(14:06):
Coming to the business saying,Hey, from ITM here to automate
your work is probably not ideal.
Business is going veryapprehensive.
A lot of fears of job protectionare out there and a lot of them
are probably unfounded, but itdoesn't matter.
They're still figures and westill need to deal with change
management.
And I choose to deal with changemanagement from the bottom up.
(14:27):
So if the business brings theideas to us and we want to do
it, not the business leader, butthe employees within the
business line, I know that theydon't need to change management
employees already happy to doit.
Maybe we can steer it, maybe wecan help them to scale it.
Maybe we can help them to thinkbigger.
But to me that's how do yourecord the value in a consistent
(14:48):
manner so you don't have an AIproject that saves millions of
dollars, and those millionsnever hit the p and l.
They never hit any financialstatements.
So those are.
Imaginary dollars.
It's fine to report it that wayif that's what the fellowship
organization wants us to do.
But most of the time they wantus to report on the real dogs
and let's report within thenumbers and the guidelines that
(15:09):
they want us to report.
Within.
Saves time save frictions makeslife easier.
And employee education.
It might be sitting within abusiness within it or somewhere
else, but to me it's one of thestaples.
If you don't have employees knowwhat to ask for, they will not
ask for the right things.
If you don't have the employeeswho know how the technology
works, they will come withunrealistic expectations.
(15:32):
Asking for the wrong thing,having realistic expectations is
never useful.
Andreas Welsch (15:37):
Wow, there was so much good information in
there.
Let me let me pick out a fewthings.
What I heard you say was, weneed to report on hard actual
savings, not just the funnymoney that we shift around
between business units but howare we actually making impact to
the business, to the P&L that'swhat our business wants and we
(15:58):
can report it and we should.
So we showed the real impact.
I love that because I think toomany times we still see the
sandbox projects or pilots.
Let's figure out what we do withAI and then maybe we roll it out
and it's actually not that easy.
But to your point, if you startwith that from the very
beginning, align it with yourbusiness strategy.
Take something that'smeasurable, that solves that
problem, right?
(16:18):
You don't even run into it.
The other part I wanted to askis, how do you get employees to
that point that they are awareof?
What can the technology actuallydo?
What is available at ourdisposal?
And could this even be somethingthat we can solve with
technology?
How do you do that?
Maxim Ioffe (16:34):
It's all about that continuous education.
I'm not a marketing person, butI've heard something talking
about pre attaches to theperson.
You need to pass thatinformation at least three times
before it ings.
I took that to heart and I triedto do it more than three times.
Every opportunity I get to talkto employees to present, it
could be a lunch and learn, itcould be five minutes.
(16:56):
During the town hall, we haveautomation minute, if you will,
where I just go in and present.
I offer it every call,everything I do.
Hey, if you want to do targetedpresentation for your employees,
happy to do it.
It's up to us who understand thetechnology and can speak speak
about it not in the IT language,but in the language that
(17:16):
business can relate to andunderstand.
Be the ambassadors of thetechnology.
It'll not happen on its own, andemployees have bombarded by
advertisement and thatadvertisement with all the right
intentions.
Create a very unrealisticpicture.
We are going to put AI in placefor your organization that will
reconcile every invoice.
(17:36):
We're going to put AI into theyour organization that is going
to take care of all therecruitment.
P two P is going to be handledby our AI agent.
How often do we see that onLinkedIn feeds?
How often do we see that duringthe presentations, equipping the
employees about it?
Talking about it?
And I, again, just right beforethis call, I was talking about
Gartner's hype curve to newemployees.
(18:00):
And we ask, say, look, when youget a new presentation about
technology.
Think where it falls into thathype.
Is it on peak of the inflatedexpectation or is it within the
trout of disillusionment?
It doesn't have to be thatterminology, but it gives you
some sort of a framework to talkabout, look, this is what is
needed.
This is how you identify thereality versus hype.
(18:22):
By the way, employee, if youbring to me something that
sounds great, but we can neverimplement it.
Nobody wins, right?
So if you know that it all logs,let's talk about it upfront and
see how we can deal with it.
There's no magic in thattechnology.
Again, a lot of folks try topretend that AI is some sort of
a magical tool that does a lotof things.
(18:43):
In a way it is.
But the reality.
Teaching the employees to checkfor that foundation is very
effective from a standpointthat, again, we're not experts.
Typically, the people whoimplement AI are not experts in
the business problem.
Yeah.
So we cannot say if thatfoundation is solid, we need
employees to help us with it.
Andreas Welsch (19:00):
That's awesome.
Maxim, there's been so manygreat pieces of information, so
many nuggets that you've sharedin the last 20, 25 minutes.
And I was wondering if you cansummarize the key three
takeaways for our audiencetoday.
Maxim Ioffe (19:13):
I wish I had your LinkedIn post from yesterday
open.
You summarized it really well.
Where the technology is goingand building 10 bots is not a
good strategy.
Three enough, 10 is not a goodstrategy.
The strategy needs to be aboutthe technology, measurable
impact of that technologies andthe strategy needs to be
something that you are willingand going to execute and include
(19:33):
at North Star.
Where do you see it going andhow do we get there from where
we are to where we wanna bewithin that strategy?
We wanna make sure that we aresupported by funding and we know
how to be funded, how to executeit.
And then the governance.
The governance is a key from astandpoint that you cannot treat
each project as a one up.
It has to be an autopilot.
(19:54):
And if it's an autopilot, thereis no friction between it and
business and everybody else.
Everybody expects and respectsthose boundaries and follows
them.
Why become easy?
You spend less time arguing,more time doing, and that's the
goal.
And the last thing I will say,maybe it's number four, is
measure twice.
Cut once.
Prep, the prep, all the things.
(20:15):
Figure out the strategy, thegovernance, everything else
before you buy the licenses,before you sign the agreement,
before you start implementing.
It makes life a lot easier and alot less mistakes are again,
getting made.
So thus would be my summary.
Andreas Welsch (20:30):
Fantastic.
Thank you so much folks.
We're at the end of the show.
Thank you so much, Maxim, forjoining us today and for sharing
your experience with us.
It was a pleasure having you onagain.
Maxim Ioffe (20:40):
My pleasure.
Thank you so much for having me.
Andreas Welsch (20:42):
Wonderful.
Popular Podcasts
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
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.