September 22, 2022 • 21 mins
In this episode, Chris Johannessen (Editor of the Journal of AI, Robotics and Workplace Automation) and Andreas Welsch discuss what to consider when setting up your Artificial Intelligence (AI) & automation Center of Excellence (CoE). Chris shares his career insights on creating CoEs and provides valuable advice for listeners considering whether to establish a CoE in their business.
Key topics:
- Determine if your business needs a CoE
- Consider organizational models and components
- Understand success criteria for CoEs
Listen to the full episode to hear how you can:
- Assess the organization’s maturity for a CoE model
- Determine your mission, vision, and roadmap
- Establish a digital presence to showcase your CoE
Watch this episode on YouTube: https://youtu.be/dWxBEivAQTI
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):
Today we'll talk about setting up your center of
excellence and who better totalk about it than someone who's
really passionate about it andwho's been doing that for a good
part of their career.
Chris Johannessen.
Hey, Chris.
Thanks for joining.
Chris Johannessen (00:15):
Hey, thank you, Andreas.
It's it's great to be here andto share a few things in 15
minutes or less with theaudience today.
Andreas Welsch (00:24):
That's awesome.
Hey, why don't you tell us alittle bit more about yourself,
who you are and what you do.
Chris Johannessen (00:29):
Sure.
Chris Johannesson.
I've been doing work with CoEs,knowledge management, digital
and technology and analytics anddata transformation for well
over 20 years now, I'm notafraid to say, you can see by
the little bit of gray hair ontop of my head here.
I've had the opportunity to getup to speed and learn this in
(00:49):
the early days of my careerwith.
A nice run at General Electricwhere I was facilitator of
knowledge management.
I even had a co o e for how todo CoEs within ge, except we
call them communities ofpractice.
And back in the day.
And then along the way, someonedecided that the word excellence
sounded better than practice anda little bit more exciting.
(01:10):
And I think I would agree.
And then from then I've hadother different stops in my
career journey buildingmarketing CoEs with clients and
internal organizations such aseBay Enterprise and some others.
And then onwards into the worldof management consultancy where
I've had some stints withorganizations such as McKinsey
(01:33):
See a partners globalConsultancies and Access group
most recently, which is more ofa boutique provider in the data
and analytics space, but helpssets the stage for automation
and AI with great data.
And when I'm not doing all that,I actually do find time for
other things.
I'm editor of the Journal of AI,robotics and Workplace
(01:56):
Automation published by HenryStewart Publications out of the
U.K.
And I've been doing that.
We just had our first full yearof four issues, so we're
celebrating Volume one.
And I do some academic advisoryas well with some universities
in the Greater Philadelphiaregion.
Andreas Welsch (02:15):
Fantastic.
Hey, really, it sounds like youare the person to talk to when
it's when it's all about CoEs.
I'm really glad that you joined.
I know we've met earlier thisyear even in person and talked
about some of these things.
So I'm really excited that we'reable to make this happen and
you're able to join.
So thanks again.
Chris Johannessen (02:36):
You're welcome
Andreas Welsch (02:37):
For those of you just joining the stream, drop a
comment in the chat if you'releading a CoE or if you are or
have been working in one, or ifyou just wanna learn more about
it and maybe join one and see ifthat is, is an organizational
model that would appeal to you.
But maybe Chris, should we playa little game and kick things
(02:57):
off?
Chris Johannessen (02:58):
Sure.
Why not?
Andreas Welsch (03:00):
Awesome.
So game time.
So this game is called In YourOwn Words.
When I hit the buzzer, thewheels will start spinning and
when they stop, you'll see asentence and I'd like you to
answer with the first thing thatcomes to mind and why, in your
(03:20):
own words.
And, to make it a little moreinteresting, you'll only have 60
seconds for your.
Are you ready for, What's theBUZZ?
Chris Johannessen (03:29):
I'm ready.
I'm getting ready player one.
Okay, perfect.
If AI were a tool in a hardwarestore, what would it be?
If it were a tool in a hardwarestore, it would be something
that could help me find tool,other tools much more easily
than my wanderings around ourlocal Lowe's and Home Depot and
(03:53):
even the the family run hardwareshop that's in my in my area.
I honestly sometimes spend waytoo much time looking for the
right size, nut, bolt screw whenI'm trying to do some household
project.
Andreas Welsch (04:06):
Fantastic.
And well within the minute.
When you said something thatcould help me find other tools,
I was envisioning it directory,but how you described it sounds
a lot more personalized and alot more specific to your needs.
Chris Johannessen (04:19):
Yeah, it could be like so we have in our
local supermarkets in the stateof Pennsylvania, one of them has
a robot that kind of rolls,rolls around the aisles, and it
could be something like thatwith a funny face on.
Talks to me, gives me clues.
Andreas Welsch (04:34):
Very true.
Yes.
And my little ones are neverquite sure whether to be excited
or to be scared when it followsthem through the aisle Anyways.
So let's jump into ourquestions.
I feel a lot of times these daysif I say automation, you say
CoE.
(04:55):
It's that well connected.
And it seems that, at least tome, it's like the go-to org
setup in a company if you wantto get started with AI and
automation.
But before going there and intothat default model, maybe let me
ask you this, when should youactually set up a CoE and when
shouldn't you?
Chris Johannessen (05:12):
That's an interesting question.
It depends on the industryvertical that your company is
in.
It depends on the size of thecompany as well.
Obviously for a very smallcompany.
Andreas might be the CEO, theone person CoE.
There's a funny meme that floatsaround the internet and LinkedIn
recently about where's all thecompany knowledge.
(05:32):
And then there's a picture ofsome person that says, I am the
knowledge.
So there's certainly some scaleconsiderations in terms of the
the current state, but also thefuture state aspirations.
Other things might be you haveto do a little bit of just like
somebody might be in terms of alittle bit of startup thinking,
(05:53):
do a little exploring and see ifsomebody's doing some automation
already within the company.
Typically you find pockets ofautomation in areas such as
maybe finance with dependingagain on a company with some
risk automation or somethinglike.
In marketing, those are areaswhere you might to see some
(06:14):
automation in play as well evenin call center operations or
other internal operations.
So it all depends on the theindustry vertical.
Certainly for the organizationsat SAP tends to work with, and
I'm sure a lot of our audienceis part of today.
Larger scale organizations startto think towards an automation
(06:35):
CoE and with that you have tohave.
Shared definition of whatautomation actually is in the
first place.
Like I said, some people mighthear the word automation and
just think, oh, reportingautomation, the analytics people
ready to do that.
You don't need a CoE if you'relooking to bring something
rather new and innovative inrobotic process automation, AI
(06:55):
is a service software platformsor even smart workflow that are
out there now that have some AIcapabilities.
There might be some workflowalready existing in the
organization, but if not, thatwould certainly be a ripe op
opportunity for bringingautomation in.
Now AI is a different animal.
(07:17):
I've heard different people saymaybe AI should just be part of
the automation CoE.
I've heard some people recentlysay AI should be part of
analytics because in order to doproper AI.
People think AI they tend todefault to statistical ways of
doing decisioning, such asmachine learning based type of
(07:39):
AI or semantic rules, engines,types of AI.
And with that, you need datascientists who build models who
then push the models intosomething and where else is a
better place in that than withthe analytics team.
Although sometimes they say no,it should be with the automation
team because the endpoint is tomake this stuff automated in
(08:02):
production, whether it'sinternally facing or customer
facing, depending on context.
Andreas Welsch (08:09):
So maybe building on that, what would you
say is like the key role of theCoE when you when you set it up
where, when you have a dedicatedteam for it?
Chris Johannessen (08:18):
First and foremost, the key role is gonna
be educating the organization.
About just what automation andAI is in that particular
context.
And with that, some of the oldermodels I mentioned, back in the
day we just did a one size fitsall CoE and you know how
(08:38):
commercials used to be, writethe ad for the Super Bowl, write
a big giant check, onecommercial fits all.
But now you have to think aboutnot only the audiences
internally, bringing a littlebit of newer thinking.
Some people might be aware ofthe idea of design thinking and
thinking about personas withinthe organization.
You're gonna be targetingautomation and AI as a value
(09:01):
driver to different people and acompany, not only their
different role levels andresponsibility, but also
different functions as wellacross, we're leader of
operations versus leader ofmarketing.
So you have to get that storyout there.
What's the value proposition?
It may start as simple as aslideshow.
(09:23):
You may do some company eventsor lunch and learns, things like
that to draw and drive someawareness or things like that.
And with that, you start toidentify, I noticed someone
dropped in the idea ofchampions.
It's always good to find achampion, someone who is either
already excited about the topic.
Could just be an executive whosaw some machine shining and
(09:45):
magical at an event and said, Ineed that too.
Or they heard competitor 1 23has it and they feel like they
have a gap as company numberfour in a vertical that they've
gotta do some catch up.
Andreas Welsch (09:56):
I know I'm a little late to taking a look at
the chat, but I see some answersto our tool question from hammer
to adjustable wrench, Swiss Armyknife.
A lot of good thoughts andsuggestions there.
So great, you set up this CoE.
They should help educate andenable parts of the organization
about what AI auto or automationor whatever other topic is.
(10:20):
And it's great, right?
As a leader, to your point,you've made that decision I need
a coE.
I understand why I would wantone.
Maybe you've been just hired toset one up.
And I feel one of the next stepsis what should that org
structure and that setup looklike?
And what does the engagementmodel with my stakeholders look
like?
So I'm curious what do you'veseen there and what you
(10:42):
recommend?
Chris Johannessen (10:43):
As with any kind of new anything resembling
a program in an organization atleast with the stakeholders, you
want to, say, you have thismission, almost like a mission
and a vision for what the CoE isall about.
Not only the materials to getpeople motivated and excited
about the rewards of automation,but also the risk as well.
(11:07):
But you've gotta be able toeducate the audience on what the
risk and rewards and talk abouta timeline and a process for how
you're gonna start bringing theCoE to life.
Like what does the early stageslook like?
What's like the middle stagesand what's the ideal end state
of what the CoE's supposed todo.
And of course, CoE's supposed tobe more about just educating,
(11:27):
getting people excited.
You've gotta have policies andprocedures going towards the
risk side, you have to becareful.
And that reminded me of yourremark earlier this week on
LinkedIn where you don't wantthe CoE to be like a cop and
stopping presses and people nolonger get excited.
But all the things in between,who are the experts CoE might
(11:48):
just start with simple as anexpert directory.
But then you could talk aboutwhat does it take to bring these
things to life in terms of aprocess evaluating the CoE.
What are some metrics for theCoE going forward?
There's certainly.
an array of differentconsiderations.
And again, it depends not onlyon the audience, but you also
have to think in another riskfactor of the especially with
(12:11):
some of the more sensitive areaswith automation lately, things
such as ethics might come up toquestion how do ethics play a
role?
Andreas Welsch (12:20):
Perfect.
So maybe taking one questionfrom Jesse.
So you're saying it's importantto have the ability to speak to
each different stakeholder inthe terms of the different
things that they want out of theCoE, so related alignment,
right, to paraphrase what yousaid.
Now I'm sure everyone is curiousto hear maybe about one example
(12:40):
where you've seen this CoE setupbe really successful and why
that has been the case?
What has made it successful?
Chris Johannessen (12:48):
We've had some recent examples or setting
up a CoE with a high technologycompany.
That was one where automationwas actually brought in to the
data and analytics CoE, as Imentioned, that could be one
example.
Because in that organizationautomation was pretty new, but
they felt data and analyticspeople that create models, were
(13:11):
just gonna have all theautomation, the intelligent
automation and AI stuff as partof the CoE, part of what helped
bring that to life in terms of.
Looking at a little bit moremacro levels, having the right
operating model in place iscertainly a consideration.
Depending on the organization,some organizations might start
(13:33):
slow and have things a bit morecentralized.
It's well known that there arethree or four different
archetypes.
Some people call them organizingmodels for operating models
where you might eventually scaletowards something where you have
a central group.
really setting all thestandards, managing all the
tools, the licenses and allsorts of other things, but
(13:53):
educating the rest of theorganization to be a bit more
self-service and help hopefullyget to the point where they may
even become, as you might hearthese days, the idea of a
citizen developer, citizenautomator.
I haven't quite heard anybodysay that they're a citizen AI
specialist yet, but you maybethree or five years down to
time.
(14:14):
timeline that it might you mightsee something like that, but
that was one where it reallytook off.
Another one was one where partof what drove the excitement was
putting some branding around theco o e almost making it almost
like a digital storefront ascertainly a component,
especially in a modern context.
Some CoEs that that don'tsucceed are weaken that regard.
(14:37):
I've seen people try to drive aCoE off of a share drive which
of course is not gonna.
Or just a simple internet pageor two.
But when you have a ratherrobust internal digital presence
that you can even put somecustom branding around and even
do some fun things, create someevents lunch and learns You'll
(14:58):
see a legacy tool and behind me,something I'd used before the
rise of mural and other toolslike Figma was I would I
actually had a a marker board onwheels and I would just wheel
it, people would laugh.
I would wheel it onto theelevator of one of the places I
was at, go up a floor and Iwould do lessons for half an
(15:18):
hour right at somebody's desk,diagramming, drawing.
If I couldn't do somethinglarger and more of a context or
a classroom.
Automation can be tricky.
In terms of success and failure.
You might have to do somebroader organizational lunch and
learns.
With one of my clients inanother part of the world, we
(15:42):
did a large scale lunch andlearn for the entire
organization to sit in to helpallay their concerns about
automation, because at thattime, there was quite a bit of
press about the threats ofautomation.
So we showed some research thatlike, look, there's more stuff
to do than there's gonna beenough people.
Lots of stuff to do.
(16:04):
People down below and in betweenis there, there's way more to do
than there's enough people tobasically handle the work.
And sometimes you have to playit safe with a little bit of a
little bit of internal educationto just kinda get people to be
like, automation's not a scarything.
It's actually gonna help youout, free you up.
Maybe you take a class cuz youcan't do things cuz you're stuck
(16:26):
doing some chore or task threedays outta the month, every
single month.
That could potentially beautomated.
Andreas Welsch (16:33):
That's a great point.
Really helping take away thatfear and show that it's not as
scary as you might think whenyou hear AI or automation,
right?
There's one other question Ihave for you and specifically
around the organizationalmodels.
Where or how do you see CoEstarting?
(16:54):
Is it typically that it'scentralized and then over time
maybe it matures in, intosomething like a federated model
where maybe the CoE looks moreafter a standardization in
governance in these type ofthings, but the individual
satellites are independent andrun the day-to-day on their own?
Or do you see companies startingalready with the federated
model?
(17:16):
Or maybe other, even othermodels than just those two that
you've seen be successful.
Chris Johannessen (17:21):
Especially when you're bringing a CoE to
life it's just good to alwayskeep that old maximum of crawl,
walk, run.
I have seen some companies tryto run right to the federated
model, which some people somepeople do see that as like the
for those of you not in, whoaren't aware, there are these
different models.
You start with a central model.
(17:43):
There might be something thatevolves towards a hub and spoke,
and then be, things become alittle bit more decentralized.
I've heard some of these thingsbe called competency centers
before you get to CoE in termsof a maturity model.
But maturity could be somethingthat's a bit more of a misnomer
that you have to watch out forbecause for some, you, someone
(18:03):
might show up and give you allthis advice.
Management consultant here,giving you a little bit of
warning about some of the otherfriends I have out there in the
world.
And maturity can sometimes beconfused with complexity where
they show you all thiscomplicated stuff.
You're like, oh, how can I getthere?
It's like you don't have toworry about getting there.
You just have to be worry abouthaving an impact and driving
(18:25):
value in your organization.
And some of these that I've seenfizzle is they try to do that,
try to run right towards theaspirations and the end goal
rather than getting some winsalong the way and kind of
building up your story andscaling up the CoE over time.
Andreas Welsch (18:42):
That's awesome.
Thanks for sharing that.
I think that's a very importantpart for anyone wondering, will,
how do I actually.
Chris Johannessen (18:49):
Another insight that I'll share there
real quick is when depending onthe organization, a CoE might be
just part of somebody's existingjob and if you don't have the
right personality in place forthat, they can fizzle cuz
they're just overworked, they'renot excited.
They see it as another thingthat you throw on top of their
list and they're just like, oh,I gotta do this.
I gotta do all this other stuff.
I got wife kids or husband kids,or, partner and kids, depending
(19:14):
on what's going on and andpersonal life, too.
Andreas Welsch (19:18):
Got it.
So maybe can you summarize thethree key points for our
audience to take away fromtoday's episode?
We talked a bit about do youeven need one?
Do you even need a CoE?
What are the organizationalmodels and when does it work
well?
What should people take awayfrom today?
Chris Johannessen (19:35):
The first thing is to make sure that you
do the right kind of assessmentbefore you even try to bring a
CoE into place.
It's good to have a some kind ofassessment to help you score the
organization, do your owninternal research.
Like I said, think of it as likea startup product market fit.
The next is if you do find anaudience and receptivity, start
to paint the picture of yourmission and your vision and your
(19:58):
roadmap for how the CoE can cometo life, and then to pilot it
and find some wins.
And last but not least, ainternal digital storefront is
usually the way to go because atleast you can have a link to
that and people can email a linkout and say, learn more about my
CoE, and send a link instead oftoying them with a big
PowerPoint deck or something.
Andreas Welsch (20:20):
Awesome.
Thanks so much for the summary.
Now, folks, we're getting closeto the end of the show.
Thank you so much for joining usand for sharing your expertise
and for those of you in theaudience for learning with us.
Thanks.
Chris Johannessen (20:32):
No, thank you for the opportunity.
There's plenty to talk abouthere.
Today we'll talk about setting up your center of
excellence and who better totalk about it than someone who's
really passionate about it andwho's been doing that for a good
part of their career.
Chris Johannessen.
Hey, Chris.
Thanks for joining.
Chris Johannessen (00:15):
Hey, thank you, Andreas.
It's it's great to be here andto share a few things in 15
minutes or less with theaudience today.
Andreas Welsch (00:24):
That's awesome.
Hey, why don't you tell us alittle bit more about yourself,
who you are and what you do.
Chris Johannessen (00:29):
Sure.
Chris Johannesson.
I've been doing work with CoEs,knowledge management, digital
and technology and analytics anddata transformation for well
over 20 years now, I'm notafraid to say, you can see by
the little bit of gray hair ontop of my head here.
I've had the opportunity to getup to speed and learn this in
(00:49):
the early days of my careerwith.
A nice run at General Electricwhere I was facilitator of
knowledge management.
I even had a co o e for how todo CoEs within ge, except we
call them communities ofpractice.
And back in the day.
And then along the way, someonedecided that the word excellence
sounded better than practice anda little bit more exciting.
(01:10):
And I think I would agree.
And then from then I've hadother different stops in my
career journey buildingmarketing CoEs with clients and
internal organizations such aseBay Enterprise and some others.
And then onwards into the worldof management consultancy where
I've had some stints withorganizations such as McKinsey
(01:33):
See a partners globalConsultancies and Access group
most recently, which is more ofa boutique provider in the data
and analytics space, but helpssets the stage for automation
and AI with great data.
And when I'm not doing all that,I actually do find time for
other things.
I'm editor of the Journal of AI,robotics and Workplace
(01:56):
Automation published by HenryStewart Publications out of the
U.K.
And I've been doing that.
We just had our first full yearof four issues, so we're
celebrating Volume one.
And I do some academic advisoryas well with some universities
in the Greater Philadelphiaregion.
Andreas Welsch (02:15):
Fantastic.
Hey, really, it sounds like youare the person to talk to when
it's when it's all about CoEs.
I'm really glad that you joined.
I know we've met earlier thisyear even in person and talked
about some of these things.
So I'm really excited that we'reable to make this happen and
you're able to join.
So thanks again.
Chris Johannessen (02:36):
You're welcome
Andreas Welsch (02:37):
For those of you just joining the stream, drop a
comment in the chat if you'releading a CoE or if you are or
have been working in one, or ifyou just wanna learn more about
it and maybe join one and see ifthat is, is an organizational
model that would appeal to you.
But maybe Chris, should we playa little game and kick things
(02:57):
off?
Chris Johannessen (02:58):
Sure.
Why not?
Andreas Welsch (03:00):
Awesome.
So game time.
So this game is called In YourOwn Words.
When I hit the buzzer, thewheels will start spinning and
when they stop, you'll see asentence and I'd like you to
answer with the first thing thatcomes to mind and why, in your
(03:20):
own words.
And, to make it a little moreinteresting, you'll only have 60
seconds for your.
Are you ready for, What's theBUZZ?
Chris Johannessen (03:29):
I'm ready.
I'm getting ready player one.
Okay, perfect.
If AI were a tool in a hardwarestore, what would it be?
If it were a tool in a hardwarestore, it would be something
that could help me find tool,other tools much more easily
than my wanderings around ourlocal Lowe's and Home Depot and
(03:53):
even the the family run hardwareshop that's in my in my area.
I honestly sometimes spend waytoo much time looking for the
right size, nut, bolt screw whenI'm trying to do some household
project.
Andreas Welsch (04:06):
Fantastic.
And well within the minute.
When you said something thatcould help me find other tools,
I was envisioning it directory,but how you described it sounds
a lot more personalized and alot more specific to your needs.
Chris Johannessen (04:19):
Yeah, it could be like so we have in our
local supermarkets in the stateof Pennsylvania, one of them has
a robot that kind of rolls,rolls around the aisles, and it
could be something like thatwith a funny face on.
Talks to me, gives me clues.
Andreas Welsch (04:34):
Very true.
Yes.
And my little ones are neverquite sure whether to be excited
or to be scared when it followsthem through the aisle Anyways.
So let's jump into ourquestions.
I feel a lot of times these daysif I say automation, you say
CoE.
(04:55):
It's that well connected.
And it seems that, at least tome, it's like the go-to org
setup in a company if you wantto get started with AI and
automation.
But before going there and intothat default model, maybe let me
ask you this, when should youactually set up a CoE and when
shouldn't you?
Chris Johannessen (05:12):
That's an interesting question.
It depends on the industryvertical that your company is
in.
It depends on the size of thecompany as well.
Obviously for a very smallcompany.
Andreas might be the CEO, theone person CoE.
There's a funny meme that floatsaround the internet and LinkedIn
recently about where's all thecompany knowledge.
(05:32):
And then there's a picture ofsome person that says, I am the
knowledge.
So there's certainly some scaleconsiderations in terms of the
the current state, but also thefuture state aspirations.
Other things might be you haveto do a little bit of just like
somebody might be in terms of alittle bit of startup thinking,
(05:53):
do a little exploring and see ifsomebody's doing some automation
already within the company.
Typically you find pockets ofautomation in areas such as
maybe finance with dependingagain on a company with some
risk automation or somethinglike.
In marketing, those are areaswhere you might to see some
(06:14):
automation in play as well evenin call center operations or
other internal operations.
So it all depends on the theindustry vertical.
Certainly for the organizationsat SAP tends to work with, and
I'm sure a lot of our audienceis part of today.
Larger scale organizations startto think towards an automation
(06:35):
CoE and with that you have tohave.
Shared definition of whatautomation actually is in the
first place.
Like I said, some people mighthear the word automation and
just think, oh, reportingautomation, the analytics people
ready to do that.
You don't need a CoE if you'relooking to bring something
rather new and innovative inrobotic process automation, AI
(06:55):
is a service software platformsor even smart workflow that are
out there now that have some AIcapabilities.
There might be some workflowalready existing in the
organization, but if not, thatwould certainly be a ripe op
opportunity for bringingautomation in.
Now AI is a different animal.
(07:17):
I've heard different people saymaybe AI should just be part of
the automation CoE.
I've heard some people recentlysay AI should be part of
analytics because in order to doproper AI.
People think AI they tend todefault to statistical ways of
doing decisioning, such asmachine learning based type of
(07:39):
AI or semantic rules, engines,types of AI.
And with that, you need datascientists who build models who
then push the models intosomething and where else is a
better place in that than withthe analytics team.
Although sometimes they say no,it should be with the automation
team because the endpoint is tomake this stuff automated in
(08:02):
production, whether it'sinternally facing or customer
facing, depending on context.
Andreas Welsch (08:09):
So maybe building on that, what would you
say is like the key role of theCoE when you when you set it up
where, when you have a dedicatedteam for it?
Chris Johannessen (08:18):
First and foremost, the key role is gonna
be educating the organization.
About just what automation andAI is in that particular
context.
And with that, some of the oldermodels I mentioned, back in the
day we just did a one size fitsall CoE and you know how
(08:38):
commercials used to be, writethe ad for the Super Bowl, write
a big giant check, onecommercial fits all.
But now you have to think aboutnot only the audiences
internally, bringing a littlebit of newer thinking.
Some people might be aware ofthe idea of design thinking and
thinking about personas withinthe organization.
You're gonna be targetingautomation and AI as a value
(09:01):
driver to different people and acompany, not only their
different role levels andresponsibility, but also
different functions as wellacross, we're leader of
operations versus leader ofmarketing.
So you have to get that storyout there.
What's the value proposition?
It may start as simple as aslideshow.
(09:23):
You may do some company eventsor lunch and learns, things like
that to draw and drive someawareness or things like that.
And with that, you start toidentify, I noticed someone
dropped in the idea ofchampions.
It's always good to find achampion, someone who is either
already excited about the topic.
Could just be an executive whosaw some machine shining and
(09:45):
magical at an event and said, Ineed that too.
Or they heard competitor 1 23has it and they feel like they
have a gap as company numberfour in a vertical that they've
gotta do some catch up.
Andreas Welsch (09:56):
I know I'm a little late to taking a look at
the chat, but I see some answersto our tool question from hammer
to adjustable wrench, Swiss Armyknife.
A lot of good thoughts andsuggestions there.
So great, you set up this CoE.
They should help educate andenable parts of the organization
about what AI auto or automationor whatever other topic is.
(10:20):
And it's great, right?
As a leader, to your point,you've made that decision I need
a coE.
I understand why I would wantone.
Maybe you've been just hired toset one up.
And I feel one of the next stepsis what should that org
structure and that setup looklike?
And what does the engagementmodel with my stakeholders look
like?
So I'm curious what do you'veseen there and what you
(10:42):
recommend?
Chris Johannessen (10:43):
As with any kind of new anything resembling
a program in an organization atleast with the stakeholders, you
want to, say, you have thismission, almost like a mission
and a vision for what the CoE isall about.
Not only the materials to getpeople motivated and excited
about the rewards of automation,but also the risk as well.
(11:07):
But you've gotta be able toeducate the audience on what the
risk and rewards and talk abouta timeline and a process for how
you're gonna start bringing theCoE to life.
Like what does the early stageslook like?
What's like the middle stagesand what's the ideal end state
of what the CoE's supposed todo.
And of course, CoE's supposed tobe more about just educating,
(11:27):
getting people excited.
You've gotta have policies andprocedures going towards the
risk side, you have to becareful.
And that reminded me of yourremark earlier this week on
LinkedIn where you don't wantthe CoE to be like a cop and
stopping presses and people nolonger get excited.
But all the things in between,who are the experts CoE might
(11:48):
just start with simple as anexpert directory.
But then you could talk aboutwhat does it take to bring these
things to life in terms of aprocess evaluating the CoE.
What are some metrics for theCoE going forward?
There's certainly.
an array of differentconsiderations.
And again, it depends not onlyon the audience, but you also
have to think in another riskfactor of the especially with
(12:11):
some of the more sensitive areaswith automation lately, things
such as ethics might come up toquestion how do ethics play a
role?
Andreas Welsch (12:20):
Perfect.
So maybe taking one questionfrom Jesse.
So you're saying it's importantto have the ability to speak to
each different stakeholder inthe terms of the different
things that they want out of theCoE, so related alignment,
right, to paraphrase what yousaid.
Now I'm sure everyone is curiousto hear maybe about one example
(12:40):
where you've seen this CoE setupbe really successful and why
that has been the case?
What has made it successful?
Chris Johannessen (12:48):
We've had some recent examples or setting
up a CoE with a high technologycompany.
That was one where automationwas actually brought in to the
data and analytics CoE, as Imentioned, that could be one
example.
Because in that organizationautomation was pretty new, but
they felt data and analyticspeople that create models, were
(13:11):
just gonna have all theautomation, the intelligent
automation and AI stuff as partof the CoE, part of what helped
bring that to life in terms of.
Looking at a little bit moremacro levels, having the right
operating model in place iscertainly a consideration.
Depending on the organization,some organizations might start
(13:33):
slow and have things a bit morecentralized.
It's well known that there arethree or four different
archetypes.
Some people call them organizingmodels for operating models
where you might eventually scaletowards something where you have
a central group.
really setting all thestandards, managing all the
tools, the licenses and allsorts of other things, but
(13:53):
educating the rest of theorganization to be a bit more
self-service and help hopefullyget to the point where they may
even become, as you might hearthese days, the idea of a
citizen developer, citizenautomator.
I haven't quite heard anybodysay that they're a citizen AI
specialist yet, but you maybethree or five years down to
time.
(14:14):
timeline that it might you mightsee something like that, but
that was one where it reallytook off.
Another one was one where partof what drove the excitement was
putting some branding around theco o e almost making it almost
like a digital storefront ascertainly a component,
especially in a modern context.
Some CoEs that that don'tsucceed are weaken that regard.
(14:37):
I've seen people try to drive aCoE off of a share drive which
of course is not gonna.
Or just a simple internet pageor two.
But when you have a ratherrobust internal digital presence
that you can even put somecustom branding around and even
do some fun things, create someevents lunch and learns You'll
(14:58):
see a legacy tool and behind me,something I'd used before the
rise of mural and other toolslike Figma was I would I
actually had a a marker board onwheels and I would just wheel
it, people would laugh.
I would wheel it onto theelevator of one of the places I
was at, go up a floor and Iwould do lessons for half an
(15:18):
hour right at somebody's desk,diagramming, drawing.
If I couldn't do somethinglarger and more of a context or
a classroom.
Automation can be tricky.
In terms of success and failure.
You might have to do somebroader organizational lunch and
learns.
With one of my clients inanother part of the world, we
(15:42):
did a large scale lunch andlearn for the entire
organization to sit in to helpallay their concerns about
automation, because at thattime, there was quite a bit of
press about the threats ofautomation.
So we showed some research thatlike, look, there's more stuff
to do than there's gonna beenough people.
Lots of stuff to do.
(16:04):
People down below and in betweenis there, there's way more to do
than there's enough people tobasically handle the work.
And sometimes you have to playit safe with a little bit of a
little bit of internal educationto just kinda get people to be
like, automation's not a scarything.
It's actually gonna help youout, free you up.
Maybe you take a class cuz youcan't do things cuz you're stuck
(16:26):
doing some chore or task threedays outta the month, every
single month.
That could potentially beautomated.
Andreas Welsch (16:33):
That's a great point.
Really helping take away thatfear and show that it's not as
scary as you might think whenyou hear AI or automation,
right?
There's one other question Ihave for you and specifically
around the organizationalmodels.
Where or how do you see CoEstarting?
(16:54):
Is it typically that it'scentralized and then over time
maybe it matures in, intosomething like a federated model
where maybe the CoE looks moreafter a standardization in
governance in these type ofthings, but the individual
satellites are independent andrun the day-to-day on their own?
Or do you see companies startingalready with the federated
model?
(17:16):
Or maybe other, even othermodels than just those two that
you've seen be successful.
Chris Johannessen (17:21):
Especially when you're bringing a CoE to
life it's just good to alwayskeep that old maximum of crawl,
walk, run.
I have seen some companies tryto run right to the federated
model, which some people somepeople do see that as like the
for those of you not in, whoaren't aware, there are these
different models.
You start with a central model.
(17:43):
There might be something thatevolves towards a hub and spoke,
and then be, things become alittle bit more decentralized.
I've heard some of these thingsbe called competency centers
before you get to CoE in termsof a maturity model.
But maturity could be somethingthat's a bit more of a misnomer
that you have to watch out forbecause for some, you, someone
(18:03):
might show up and give you allthis advice.
Management consultant here,giving you a little bit of
warning about some of the otherfriends I have out there in the
world.
And maturity can sometimes beconfused with complexity where
they show you all thiscomplicated stuff.
You're like, oh, how can I getthere?
It's like you don't have toworry about getting there.
You just have to be worry abouthaving an impact and driving
(18:25):
value in your organization.
And some of these that I've seenfizzle is they try to do that,
try to run right towards theaspirations and the end goal
rather than getting some winsalong the way and kind of
building up your story andscaling up the CoE over time.
Andreas Welsch (18:42):
That's awesome.
Thanks for sharing that.
I think that's a very importantpart for anyone wondering, will,
how do I actually.
Chris Johannessen (18:49):
Another insight that I'll share there
real quick is when depending onthe organization, a CoE might be
just part of somebody's existingjob and if you don't have the
right personality in place forthat, they can fizzle cuz
they're just overworked, they'renot excited.
They see it as another thingthat you throw on top of their
list and they're just like, oh,I gotta do this.
I gotta do all this other stuff.
I got wife kids or husband kids,or, partner and kids, depending
(19:14):
on what's going on and andpersonal life, too.
Andreas Welsch (19:18):
Got it.
So maybe can you summarize thethree key points for our
audience to take away fromtoday's episode?
We talked a bit about do youeven need one?
Do you even need a CoE?
What are the organizationalmodels and when does it work
well?
What should people take awayfrom today?
Chris Johannessen (19:35):
The first thing is to make sure that you
do the right kind of assessmentbefore you even try to bring a
CoE into place.
It's good to have a some kind ofassessment to help you score the
organization, do your owninternal research.
Like I said, think of it as likea startup product market fit.
The next is if you do find anaudience and receptivity, start
to paint the picture of yourmission and your vision and your
(19:58):
roadmap for how the CoE can cometo life, and then to pilot it
and find some wins.
And last but not least, ainternal digital storefront is
usually the way to go because atleast you can have a link to
that and people can email a linkout and say, learn more about my
CoE, and send a link instead oftoying them with a big
PowerPoint deck or something.
Andreas Welsch (20:20):
Awesome.
Thanks so much for the summary.
Now, folks, we're getting closeto the end of the show.
Thank you so much for joining usand for sharing your expertise
and for those of you in theaudience for learning with us.
Thanks.
Chris Johannessen (20:32):
No, thank you for the opportunity.
There's plenty to talk abouthere.
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.