Hello and welcome to ProcessTransformers, the podcast that talks
about business transformation at theintersection of processes and AI.
For those of you who havelistened before, welcome back.
And if you're new to theshow, thanks for tuning in.
My name is Lukas Egger.
I'm the head of innovation atSAP Signavio, and I'll be your
host for today's episode titled" Harnessing the Power of AI to
Transform Corporate Innovation".
And I'm extremely happy to introducetoday's guest, Sean Sheppard, serial
entrepreneur, and among otherthings, managing partner at U+.
So good to have you.
Hello, Sean.
Hi Lukas, glad to be here.
Wonderful.
Sean, you and U+ are talkinga lot about incorporating AI,
especially for innovation.
Now, everybody is like this mad dashto adopt AI and make it work, but
most of the time it seems aroundefficiency plays or automating stuff.
Innovation is not typicallythe first thing that comes up.
Why is innovation the perfect placeto adopt AI and what is the upside
for a corporate environment to do so?
well I would start with kindof reframing the premise.
I think innovation is the byproductof harnessing and leveraging
a new technology like AI.
So it doesn't necessarily haveto be within the innovation
group or innovation department.
by leveraging and implementing AIto improve productivity or decrease
costs or help you be smarter andadvanced the human condition, generally
speaking, is de facto innovation.
the reason we selfishly think innovation'sa great place to do it is because a lot
of the work that's done is very manual.
It's very labor intensive.
It requires a tremendousamount of brain power.
Which is also very expensive.
and so I kind of think about it likethe Industrial Revolution replaced
muscle power with mechanical power.
the AI revolution is replacingbrain power with computer power.
and I think it's going to do itto a pretty significant degree,
and we're already seeing it happen.
I don't know if you saw Elon Musk'stalk at the Milken Institute, but he
talked about, A future where 99 percentthe brainpower will be artificial and
only 1 percent will be organic or,biological, in that future, right?
So how do you stay relevant?
How do you leverage it tothe best of your abilities?
innovation is a great place to doit because, A, we've been, building
corporate ventures for largeorganizations like yours for 15 years.
We've built over 130 new ventures andtypically the start of that process
is trying to fill the top of yourfunnel with a bunch of ideas, right?
Where should I play and what should Ido based on who I am, who my customers
are, what we're good at, what ourcore assets and capabilities are.
that process requires a lot of veryexpensive people who are very smart
and a lot of digital ditch digging.
And so we took an approach of how canwe leverage AI automate ourselves out of
business, essentially, or out of a portionof our business, because if we don't
cannibalize ourselves, someone else will.
And so we have been able to take what usedto take six months and cost hundreds of
thousands of dollars down to literallyhours and thousands of dollars to start
the process, to catalyze the creativityand the optionality, of where and how we
should play and what we should be lookingat and constructing and framing problem
statements, markets, opportunities,so potential solutions so that we can
better prepare ourselves and our clients.
to go to market and start testingthese concepts and ideas to
figure out where traction exists.
And we believe that AI will createpredictive company building in the not too
distant future, to predict unit economics,to predict problems and solutions and
needs in a way that you can go rightto the validation stage, and determine
whether or not there's a there there.
and, it's really disrupting theentire professional services
and management consulting space.
I now make the joke that in the worldof AI, if you want to be a millionaire
in management consulting, you needto start out as a multi millionaire.
Because it's not, it's notgoing to be what it once was.
I just got off a call with a clientwho for the first time saw our
platform and said, Oh my God, you'regoing to kill management consulting.
not entirely.
I think what we're learning isyou still need a trusted source
at the head of these things.
You still need a subject matter expert.
You still need a throat tochoke if things don't go well.
But how we deliver our servicesis rapidly, rapidly changing.
And now it's a platform withpeople instead of just people.
and what we're trying to do isenable not just ourselves, but our
clients to be better, smarter, faster.
I think about it in terms ofwhat I call The 90 80 rule.
Leveraging AI, especially in digitalbusinesses, with transformation
or otherwise, I can get you 90%.
I can do it 90 percent faster andcheaper and get you 80 percent
of the way there in the process.
So now the focus becomes on that last 20.
Where's the nuance?
Where's the insights?
Where is the seventh why?
How do we get that last mile?
how do we complete thatlast mile create value?
that's where we already are in justsix or seven months of doing this.
and I can't imagine, Imean, I can't imagine.
I'm seeing a world three, five yearsdown the road when corporates get past
the risk concerns and, recognize thevalue of it where, it's making all of
us 40 to 60 percent more productive.
There's so many things I desperately wantto touch upon because you have offered
so many interesting ideas right there.
I think we should startmaybe with two things.
First, how do you see an innovationprocess typically pan out in the past?
And how will it work in the futureif we incorporate AI correctly?
I think that maybe is also a level setbecause, often innovation is one word, but
how it is practiced and how it is livedis very different in different companies.
And somebody might recognize it asinnovation where somebody else will say,
well, that's just basic product discovery.
That is, not what we understand.
Under this umbrella term.
And the second thing is I reallylove what you said about replacing
brain power with compute power.
But it begs the question that itis something that can be automated.
Whereas I think a lot of the peoplein the field have this idea about
innovation, not as a manual or tritetask that can be automated, but more
of an art, right, that is not easilycaptured with something that is computer
or any automating kind of entity.
So, I would love to touch on thepast and the future and then also in
what way can AI help and maybe itis an art or maybe it's not and it's
more of a science or maybe there'sjust a lot of overhead that we can
automate to maybe go in there and clear
that one up.
Yeah.
So let's talk about the pastversus the future, because I
think the future is already here.
It's just not evenly distributed.
But, what essentially, AI hasdone, at least in our experiences
in our application has.
dramatically compressed the front endof the innovation process and moved
the bottleneck from what should wedo and where should we play to, all
right, what are we ready to test?
and so it's compressed what weused to sell as a perfect example.
It was a professional service.
It would take five to six months tohelp you come, select a space, come up
with some business model concepts, buildfully enriched idea cases around those,
and then start testing and validating.
and now we can literally do that ina matter of hours, get ready to test.
So, from questions to answers andinsights that are testable can now
be done in a matter of hours whenit used to be months and hundreds
and thousands of person hours.
That was a 2, 500 manhour, let's call it scope.
that we used to provideas a professional service.
Now we can literally do it in hours.
So it's shifting the bottleneckfurther down the path, so that you
can compress the time to adoption.
Because to me, innovation isn'tinnovation without adoption.
And that's ultimatelywhere you're trying to get.
can you get to a happy cohort of usersor customers, who will tell the world
they're better off with you than withoutyou, and why, around a specific use
case that's predictable, repeatable, andeither profitable or more ,efficient.
that's the future that's already here.
So what used to take nine monthsto a year, year and a half,
depending on the dependency of theclient, is now taking six months.
and it's driving massive efficiency.
So what does that do?
That allows you to fill the top ofyour funnel with more ideas and then
test and validate down to de riskingmore ideas that have a better chance
of getting the outcomes that you want.
so that's the first part.
the second part is, the dynamicof art and science collaborating
together to create real valuein the world isn't going away
.I think a lot of people assume,especially if they come from the area
of let's say de risking or productinnovation, that their work is more
of an art than a mechanical process.
Right?
And I think this is not only truefor people's, working in whatever
quote unquote innovation area.
But I think every task that youwill go out and say, Hey, we have
AI and we can either replace or.
massively decrease your workload.
The first thing people willsay is, hold on a second.
This is not a boring, trite task.
This is an art.
And I don't believe that AIcan actually do the work.
right, but what is happening is thisgives the artist more opportunity
to focus on what's most important.
based on automating a significantportion of the science.
And so how do you do that now?
In practical terms we callit becoming AI natives.
And AI native means you paira subject matter expert with
a technologist or a developer.
to solve problems using AI up to thepoint to where now that subject matter
expert can apply our experience,insights, subject matter expertise,
decision making into how do youtake action based on that science.
because all of that science wasdone by a lot of scientists and now
the AI is replacing a lot of thatand just making it more accurate.
and driving massive, massive efficiencies.
So, you're still going to have the artand the science, it's just I think there
will be more emphasis on the artists.
and how they choose to use that scienceto advance the human condition or
value.
Hm.
are just the same conversationsyou have with potential customers.
I guess what I'm curiousabout is there's tons of.
ideas out there with people advocatingfor them and saying you have to try
this, it will revolutionize whatever,like blockchain there, IoT over
there, what are the typical concernsand what has been your experience
in addressing those concerns?
So, for instance, I could imagine thatcompanies are worried that if they're
using something of a standard service,that it diminishes the differentiating
aspects of their ideas, that they wouldget the same ideas as others, or that it
might overwhelm their capacity for change.
Or I don't know, what are the realconcerns that you see in the wild
with your customers and what has beenthe feedback and the experiences in
addressing them or working with them.
Well, in any period of significantchange, The first, concern everyone
has, whether they say it or not, iswe fear what we don't understand.
So they immediately gotowards the risk mindset.
what are the securityconcerns around this?
What are the IP concerns around this?
What are the data integrityrisks associated with it?
Which are really important questions.
The is the most important element.
You've got to have good data.
It's garbage in garbage out otherwise.
And so how do you QAand validate that data?
That's why it's not a pure service,it's a managed service where you still
have those subject matter experts andthose prompt engineers and analysts
that are focused on making sure thatthe quality of what you're white
listing is good, and then your blacklisting the things that are not good.
So those are the primary concerns.
The next one is legislative innature, or regulatory in nature.
It's AI policy.
Not just your own internal corporatepolicy, but the policy at the federal,
state, and local levels of governments,and or at the industry level.
and how those things are impacting people.
It's so funny, we are now building anAI policy tracker as a separate spin
out, because that's been such a massivedemand from our current customers, in
partnership with UC Berkeley, who'sprovided us with all of the vetted
data sources, over 1, 600, I believe.
that we can now track and manageand deliver AI policy information
to clients in real time so theyknow how to make good choices.
So those are the, number one concerns.
Then of course there's, it'sgoing to take my job concern.
and I don't think that's theright way to think about it.
I think the way to think aboutit is it's going to make you more
productive and it's going to allowyou to focus on higher value things.
Just like the artist and the scientist.
and then the next oneis to your point, where,
you know, there are biases inorganizations and with humans
generally about what they thinkis a good idea, what they don't.
And of course, there's, political,ramifications and, influences and
there's power influences and there'sother things that drive behavior of
the people who are trying to supportan organization through innovation.
And their biggest frustration asa shared service is I don't have a
standardized way of intaking these ideas.
vetting these ideas and thendelivering in a safe way, why we
should or should not pursue this idea.
And a tool like what we'vebuilt allows you to address
all those concerns at once.
You have a standard approachto collect ideas at scale.
You have a, quick dirty way of,I would say quick and clean, way
of processing and assessing andevaluating and understanding those
through exploration and research.
and then as a result of that, youcan then deliver back in an objective
way should we test this or not?
And if we do test it, then, the customeror the user inside the organization is
going to tell us if it's any good or not.
And that's that last 20percent that we talked about.
Getting you 80 percent of the way there.
The method or the process you'redescribing sounds, I don't want to
say generic, but it sounds like it'svery applicable not only to innovation.
Do you feel like that innovation becauseit was touted or used to be more of an
art that, it's more ripe for disruption?
Or do you see that what you'recurrently testing out will also be
equally applicable to other domainsof corporate life cycle or activities?
Yes, we first focused on externalinnovation for obvious reasons.
That was our background and experience.
But very quickly, we realized people wereusing it for internal challenges as well.
Because the approach is the same.
we want new ideas.
We want to be better.
We want to advance.
We want to improve.
We want to do this.
We want to do that.
Where do we start?
And then how do we evaluatewhat should be done?
And then how do we test what should bedone to ensure that it's going to give
us the outcomes that we're looking for?
So you still have that rapid iterativelearning process that happens
anytime anything new occurs, right?
and so the idea of the AI is really justto get you to a place to where now you're
fully prepared with a well constructedhypothesis that you can test with humans.
And that's the advancement thatwe've been able to make rapidly
because that's the most obvious part.
When you're doing desk research,bench research, intelligence
research, those kinds of things, AIis a fantastic use case for that.
And of course, it's not100 percent perfect.
It's going to get you 80 percentof the way there, but it's going
to do an hour as opposed to months.
I want to go back to something yousaid before about the data integrity
and the availability of high qualitydata is of paramount concern.
Mm hmm.
I think everybody agrees, right?
Everybody wants good data and more of it.
There has never been an issuewith that, but what is good
data is very hard to know.
just like when we go backto the analogy of art.
You have to see a lot of art tohave, you know, an educated opinion.
Yeah.
to recognize art what it is.
Yeah, but you see more data fromdifferent companies, with different,
I guess, histories and contingenciesand, all that played into it.
I would love to hear your take.
What are the truths about good data ormaybe also the unintuitive lessons, like
what would empower someone to, whetherthey work with an external consultancy,
with AI, or even with, somethinginternally, what do they need to have data
wise and quality wise to be set up well?
Well, so I think the first thing isyou should always have the attitude or
approach of trust but verify, right?
and that's why we have analysts focused onthe QA ing of this information and working
with the client every day to let themtell us if they think this is good data.
Is it from a credible source?
Whether it's your internal data,it's public data, or it's data that's
being, delivered through subscriptionsto certain data sources or even
from our own internal library wherewe're training, on a regular basis
and doing industry level things.
So you've got to start with trust,but verify, Don't just take it for
what it is on its face until you'velearned how to trust it over time.
you do need to have a mechanismin place for, ensuring
quality.
When you say trust, but verify,is that for you a guided process
through a product or a process?
Or how would that pan out?
What does that mean for someone whosays, Oh, we have one petabyte of data.
We're in a perfect position.
I've verified it.
It's one petabyte of data.
We're good.
well, I think that's where you needto have a human in the loop, right?
There will still need to be ahuman in the loop or a set of
humans, the artists if you will.
that have the insights, the knowledge,the lived experiences, the subject matter
expertise, the domain information, andall of the other nuances around the world
in which they operate and move through.
they need to be, acutely aware, of,those outputs and make sure that
they're coming from the right place.
and that it's good information.
And one of the beauties of whatAI can do for you is it can pull
from a lot of different sources sothat you have basis of comparison.
and you can benchmark against work youmay have already done done manually
. for example, and we see that a lot.
as far as, good data structures, look,the end of the day, it's, if you can
crawl it, we can source it, and we cansynthesize and summarize it and give you
an opportunity to digest and then makegood actionable decisions based on that.
that's where we are today.
The future looks like visualrepresentations of that.
A lot of people are asking for, okay,now how do I take all this text based
summaries and insights and createvisual representations that tell
the right compelling story So thatI can get my organization to buy in.
and that's coming, but we're stillworking on the datasets and we'll
always be working on those datasetsbecause that market is changing rapidly,
whether it's Anthropic, of course,or, perplexity pro or open AI, or.
Co pilot, you name it, right?
so you always have to have a mind on that.
and then the results and the sources ofthose results are absolutely critical,
which is why it's critical to make surethat you have, sort of a data supply
chain in place that allows you to trackand trace source of that information and
determine whether or not it's, credible.
That was really impactful as a statementyou said in a way you were in the business
of killing consultants or the businessmodel of management consultancies.
Now, it happens to be the case thatAccenture had 600 millions in booking
for a generative AI related I thinkit's their biggest single line item
for them, and they're on track to makemore than maybe 2 billion this year.
So where's the discrepancy, right?
It seems to be like the hottestand the most flourishing part for
consultancies, gen AI, but yet,you just said this will be killed.
So how do we make sense of that?
Well, I think what you're just seeingis the transition and they're smart and
everybody pivoting towards AI strategyand implementation work because that's
what companies are doing right now.
almost every client we work withhas some sort of quote unquote AI
working group that they put togetherto try and figure out where and how
we should be implementing this stuff.
And so, for the near term, whateverthat looks like, five years, 10
years, that work will be there.
but they're also going to be able todo it much more efficiently with a lot
less, people in their organization.
And so they're going to get leanerand and there will be fewer jobs
and there'll be more output andthere will be pricing changes and
adjustments based on those things.
But over time, that's going to wanebecause if you really are using
AI the way you should be using it.
It should enabling you to dothe work of those consultants.
overwhelming majority of them.
As I said, it's not goingto be 100 percent gone.
But you will have smart, trustedsources leading these things for you.
And C suites will always demand that.
It's their nature.
They want that certainty.
And they want that human experiencedriving this stuff for them.
So, you're just going to doit with a lot less people.
80 percent of it'sgoing to go away, right?
20 percent will remain and work inthat capacity, but they'll be doing it
powered by AI and helping enable their
clients to be powered by AI.
that sounds very plausible, right?
if that's true, what will bethe strategy for companies out
there in what should they invest?
in terms of what can be thedifferentiating competitive
advantage of a company.
And also, since a lot of people willbe in one way or the other affected
by it and turning into artists orhaving to change the core of the work.
How do you see that play out?
what are the right things tosupport or do, especially since we
had other guests on this podcastpointing out that the change.
management process will be ofthe most decisive factors in
turning this into value not.
So your take
on it,
I think from a talent perspective, ifyou want to be relevant in the age of AI,
you, need to become more of a generalist,
and you need to understand how towork from the problem backwards
to leverage things like that.
That's why I talked aboutpairing subject matter experts
with, technical experts, right?
The artists and the scientists.
so to upskill and reskill yourworkforce to be more generalist in
nature is going to be super important.
then, you know, where do you start, right?
you start by where can we createbusiness value with applicable use
cases today, regardless of whether ornot you need to build it or it already
exists in the market and you can buy it.
I'm going to do thethings that either create.
top line, revenue, opportunities, growthe top line or, grow the bottom line.
so make money, save money, create andmaintain a competitive advantage and
stay out of prison as I like to say.
Risk and liability.
and there are very applicable use casesright now across the board using existing
tools that are already in the market.
And then putting your thin veneeron top of them to create that value.
and it's surprisingly more.
Surprisingly, I think to me,probably having been through
so many, look, I've been in thetech space since 96, right, 97.
So, I've been through all thesetransitions from the internet to
com from on prem to SAS, to now AI.
and this one to me has been byfar the easiest one to implement.
because many of the tools exist, tobe able to solve meaningful problems
on a use case level very quickly.
So that's what I'm seeing.
Or scale ups.
Do you see that, this appliesnow equally to smaller companies
and bigger corporations,
It does.
look, at the end of the day, all thisstuff it's not B2C or product or
market or industry or sector specific.
It's what I call H2H.
It's human to human, right?
until bots are buying on your behalffrom other bots who are selling to
them, that's not going to change.
we still want that human experience.
we still want those relationships.
We still need that trust andcertainty, and reliability.
And that's only going to comethrough human experience.
So this whole thing is H2H.
It doesn't matter if it's forinternal use or external use.
And I don't see that changing.
That's a beautiful way of puttingit, but not everything is easy.
So we always like to ask one questionof all the processes you see out
there and the impediments, like if youcould transform or change one of those
processes, which one would it be why?
well, so I'm a big fan of AI for,personal and professional development.
in fact, I've invested on theboard of a company called,
Primer AI, which is an AI.
Lifelong Learning Companion forPersonal Professional Development.
It has a wide variety of applications.
Obviously, it's K 12,universities, and individually.
what it does to help put aperson on the right path to
learn and grow is remarkable.
and I'm very excited about that future.
Is it going to replace teachers?
No.
Is it going to helpteachers be better teachers?
Yes.
Is it going to helpstudents be better students?
Yes.
I'm a big fan of the co pilotapproach towards the use of AI.
Everybody kind of jumps to theend and thinks about full agency
and gets scared by the media andthe hype and everything else.
We're nowhere near that.
There are certain elements that you canprovide full agency to, but it's very low
risk, high volume, task oriented things.
but the things that you and I cantake along with us, just like we
take along our mobile phone, thathelp make our lives better, you know,
that's what I'm most excited about
Well, that's a very upliftingand beautiful message.
And I would say also theperfect message to end on.
Thank you so much, Sean, forbeing part of the podcast today.
Good luck uh, with usheringin Gen AI for all aspects of
corporate, not just innovation.
And with that, thanks for listening in toanother episode of Process Transformers.
Thank you, Lucas.
I only wish the audience could be apart of all of our other conversations
that are lengthy and varied, andI always enjoy talking with you,
so thanks again for having me.
And with that, thanks somuch for listening to another
episode of Process Transformers.
This podcast is brought toyou by the dedicated efforts
and hard work of our team.
So our heartfelt thank you goesout to Beyza Kartal, Dharmik
Rathod, and Reagan Nyandoro.
If you have any questions or comments,email us at processtransformers at sap.
com.
And until next time, foranother transformative
about business transformation at theintersection of processes and AI.
For those of you who havelistened before, welcome back.
And if you're new to theshow, thanks for tuning in.
My name is Lukas Egger.
I'm the head of innovation atSAP Signavio, and I'll be your
host for today's episode titled" Harnessing the Power of AI to
Transform Corporate Innovation".
And I'm extremely happy to introducetoday's guest, Sean Sheppard, serial
entrepreneur, and among otherthings, managing partner at U+.
So good to have you.
Hello, Sean.
Hi Lukas, glad to be here.
Wonderful.
Sean, you and U+ are talkinga lot about incorporating AI,
especially for innovation.
Now, everybody is like this mad dashto adopt AI and make it work, but
most of the time it seems aroundefficiency plays or automating stuff.
Innovation is not typicallythe first thing that comes up.
Why is innovation the perfect placeto adopt AI and what is the upside
for a corporate environment to do so?
well I would start with kindof reframing the premise.
I think innovation is the byproductof harnessing and leveraging
a new technology like AI.
So it doesn't necessarily haveto be within the innovation
group or innovation department.
by leveraging and implementing AIto improve productivity or decrease
costs or help you be smarter andadvanced the human condition, generally
speaking, is de facto innovation.
the reason we selfishly think innovation'sa great place to do it is because a lot
of the work that's done is very manual.
It's very labor intensive.
It requires a tremendousamount of brain power.
Which is also very expensive.
and so I kind of think about it likethe Industrial Revolution replaced
muscle power with mechanical power.
the AI revolution is replacingbrain power with computer power.
and I think it's going to do itto a pretty significant degree,
and we're already seeing it happen.
I don't know if you saw Elon Musk'stalk at the Milken Institute, but he
talked about, A future where 99 percentthe brainpower will be artificial and
only 1 percent will be organic or,biological, in that future, right?
So how do you stay relevant?
How do you leverage it tothe best of your abilities?
innovation is a great place to doit because, A, we've been, building
corporate ventures for largeorganizations like yours for 15 years.
We've built over 130 new ventures andtypically the start of that process
is trying to fill the top of yourfunnel with a bunch of ideas, right?
Where should I play and what should Ido based on who I am, who my customers
are, what we're good at, what ourcore assets and capabilities are.
that process requires a lot of veryexpensive people who are very smart
and a lot of digital ditch digging.
And so we took an approach of how canwe leverage AI automate ourselves out of
business, essentially, or out of a portionof our business, because if we don't
cannibalize ourselves, someone else will.
And so we have been able to take what usedto take six months and cost hundreds of
thousands of dollars down to literallyhours and thousands of dollars to start
the process, to catalyze the creativityand the optionality, of where and how we
should play and what we should be lookingat and constructing and framing problem
statements, markets, opportunities,so potential solutions so that we can
better prepare ourselves and our clients.
to go to market and start testingthese concepts and ideas to
figure out where traction exists.
And we believe that AI will createpredictive company building in the not too
distant future, to predict unit economics,to predict problems and solutions and
needs in a way that you can go rightto the validation stage, and determine
whether or not there's a there there.
and, it's really disrupting theentire professional services
and management consulting space.
I now make the joke that in the worldof AI, if you want to be a millionaire
in management consulting, you needto start out as a multi millionaire.
Because it's not, it's notgoing to be what it once was.
I just got off a call with a clientwho for the first time saw our
platform and said, Oh my God, you'regoing to kill management consulting.
not entirely.
I think what we're learning isyou still need a trusted source
at the head of these things.
You still need a subject matter expert.
You still need a throat tochoke if things don't go well.
But how we deliver our servicesis rapidly, rapidly changing.
And now it's a platform withpeople instead of just people.
and what we're trying to do isenable not just ourselves, but our
clients to be better, smarter, faster.
I think about it in terms ofwhat I call The 90 80 rule.
Leveraging AI, especially in digitalbusinesses, with transformation
or otherwise, I can get you 90%.
I can do it 90 percent faster andcheaper and get you 80 percent
of the way there in the process.
So now the focus becomes on that last 20.
Where's the nuance?
Where's the insights?
Where is the seventh why?
How do we get that last mile?
how do we complete thatlast mile create value?
that's where we already are in justsix or seven months of doing this.
and I can't imagine, Imean, I can't imagine.
I'm seeing a world three, five yearsdown the road when corporates get past
the risk concerns and, recognize thevalue of it where, it's making all of
us 40 to 60 percent more productive.
There's so many things I desperately wantto touch upon because you have offered
so many interesting ideas right there.
I think we should startmaybe with two things.
First, how do you see an innovationprocess typically pan out in the past?
And how will it work in the futureif we incorporate AI correctly?
I think that maybe is also a level setbecause, often innovation is one word, but
how it is practiced and how it is livedis very different in different companies.
And somebody might recognize it asinnovation where somebody else will say,
well, that's just basic product discovery.
That is, not what we understand.
Under this umbrella term.
And the second thing is I reallylove what you said about replacing
brain power with compute power.
But it begs the question that itis something that can be automated.
Whereas I think a lot of the peoplein the field have this idea about
innovation, not as a manual or tritetask that can be automated, but more
of an art, right, that is not easilycaptured with something that is computer
or any automating kind of entity.
So, I would love to touch on thepast and the future and then also in
what way can AI help and maybe itis an art or maybe it's not and it's
more of a science or maybe there'sjust a lot of overhead that we can
automate to maybe go in there and clear
that one up.
Yeah.
So let's talk about the pastversus the future, because I
think the future is already here.
It's just not evenly distributed.
But, what essentially, AI hasdone, at least in our experiences
in our application has.
dramatically compressed the front endof the innovation process and moved
the bottleneck from what should wedo and where should we play to, all
right, what are we ready to test?
and so it's compressed what weused to sell as a perfect example.
It was a professional service.
It would take five to six months tohelp you come, select a space, come up
with some business model concepts, buildfully enriched idea cases around those,
and then start testing and validating.
and now we can literally do that ina matter of hours, get ready to test.
So, from questions to answers andinsights that are testable can now
be done in a matter of hours whenit used to be months and hundreds
and thousands of person hours.
That was a 2, 500 manhour, let's call it scope.
that we used to provideas a professional service.
Now we can literally do it in hours.
So it's shifting the bottleneckfurther down the path, so that you
can compress the time to adoption.
Because to me, innovation isn'tinnovation without adoption.
And that's ultimatelywhere you're trying to get.
can you get to a happy cohort of usersor customers, who will tell the world
they're better off with you than withoutyou, and why, around a specific use
case that's predictable, repeatable, andeither profitable or more ,efficient.
that's the future that's already here.
So what used to take nine monthsto a year, year and a half,
depending on the dependency of theclient, is now taking six months.
and it's driving massive efficiency.
So what does that do?
That allows you to fill the top ofyour funnel with more ideas and then
test and validate down to de riskingmore ideas that have a better chance
of getting the outcomes that you want.
so that's the first part.
the second part is, the dynamicof art and science collaborating
together to create real valuein the world isn't going away
.I think a lot of people assume,especially if they come from the area
of let's say de risking or productinnovation, that their work is more
of an art than a mechanical process.
Right?
And I think this is not only truefor people's, working in whatever
quote unquote innovation area.
But I think every task that youwill go out and say, Hey, we have
AI and we can either replace or.
massively decrease your workload.
The first thing people willsay is, hold on a second.
This is not a boring, trite task.
This is an art.
And I don't believe that AIcan actually do the work.
right, but what is happening is thisgives the artist more opportunity
to focus on what's most important.
based on automating a significantportion of the science.
And so how do you do that now?
In practical terms we callit becoming AI natives.
And AI native means you paira subject matter expert with
a technologist or a developer.
to solve problems using AI up to thepoint to where now that subject matter
expert can apply our experience,insights, subject matter expertise,
decision making into how do youtake action based on that science.
because all of that science wasdone by a lot of scientists and now
the AI is replacing a lot of thatand just making it more accurate.
and driving massive, massive efficiencies.
So, you're still going to have the artand the science, it's just I think there
will be more emphasis on the artists.
and how they choose to use that scienceto advance the human condition or
value.
Hm.
are just the same conversationsyou have with potential customers.
I guess what I'm curiousabout is there's tons of.
ideas out there with people advocatingfor them and saying you have to try
this, it will revolutionize whatever,like blockchain there, IoT over
there, what are the typical concernsand what has been your experience
in addressing those concerns?
So, for instance, I could imagine thatcompanies are worried that if they're
using something of a standard service,that it diminishes the differentiating
aspects of their ideas, that they wouldget the same ideas as others, or that it
might overwhelm their capacity for change.
Or I don't know, what are the realconcerns that you see in the wild
with your customers and what has beenthe feedback and the experiences in
addressing them or working with them.
Well, in any period of significantchange, The first, concern everyone
has, whether they say it or not, iswe fear what we don't understand.
So they immediately gotowards the risk mindset.
what are the securityconcerns around this?
What are the IP concerns around this?
What are the data integrityrisks associated with it?
Which are really important questions.
The is the most important element.
You've got to have good data.
It's garbage in garbage out otherwise.
And so how do you QAand validate that data?
That's why it's not a pure service,it's a managed service where you still
have those subject matter experts andthose prompt engineers and analysts
that are focused on making sure thatthe quality of what you're white
listing is good, and then your blacklisting the things that are not good.
So those are the primary concerns.
The next one is legislative innature, or regulatory in nature.
It's AI policy.
Not just your own internal corporatepolicy, but the policy at the federal,
state, and local levels of governments,and or at the industry level.
and how those things are impacting people.
It's so funny, we are now building anAI policy tracker as a separate spin
out, because that's been such a massivedemand from our current customers, in
partnership with UC Berkeley, who'sprovided us with all of the vetted
data sources, over 1, 600, I believe.
that we can now track and manageand deliver AI policy information
to clients in real time so theyknow how to make good choices.
So those are the, number one concerns.
Then of course there's, it'sgoing to take my job concern.
and I don't think that's theright way to think about it.
I think the way to think aboutit is it's going to make you more
productive and it's going to allowyou to focus on higher value things.
Just like the artist and the scientist.
and then the next oneis to your point, where,
you know, there are biases inorganizations and with humans
generally about what they thinkis a good idea, what they don't.
And of course, there's, political,ramifications and, influences and
there's power influences and there'sother things that drive behavior of
the people who are trying to supportan organization through innovation.
And their biggest frustration asa shared service is I don't have a
standardized way of intaking these ideas.
vetting these ideas and thendelivering in a safe way, why we
should or should not pursue this idea.
And a tool like what we'vebuilt allows you to address
all those concerns at once.
You have a standard approachto collect ideas at scale.
You have a, quick dirty way of,I would say quick and clean, way
of processing and assessing andevaluating and understanding those
through exploration and research.
and then as a result of that, youcan then deliver back in an objective
way should we test this or not?
And if we do test it, then, the customeror the user inside the organization is
going to tell us if it's any good or not.
And that's that last 20percent that we talked about.
Getting you 80 percent of the way there.
The method or the process you'redescribing sounds, I don't want to
say generic, but it sounds like it'svery applicable not only to innovation.
Do you feel like that innovation becauseit was touted or used to be more of an
art that, it's more ripe for disruption?
Or do you see that what you'recurrently testing out will also be
equally applicable to other domainsof corporate life cycle or activities?
Yes, we first focused on externalinnovation for obvious reasons.
That was our background and experience.
But very quickly, we realized people wereusing it for internal challenges as well.
Because the approach is the same.
we want new ideas.
We want to be better.
We want to advance.
We want to improve.
We want to do this.
We want to do that.
Where do we start?
And then how do we evaluatewhat should be done?
And then how do we test what should bedone to ensure that it's going to give
us the outcomes that we're looking for?
So you still have that rapid iterativelearning process that happens
anytime anything new occurs, right?
and so the idea of the AI is really justto get you to a place to where now you're
fully prepared with a well constructedhypothesis that you can test with humans.
And that's the advancement thatwe've been able to make rapidly
because that's the most obvious part.
When you're doing desk research,bench research, intelligence
research, those kinds of things, AIis a fantastic use case for that.
And of course, it's not100 percent perfect.
It's going to get you 80 percentof the way there, but it's going
to do an hour as opposed to months.
I want to go back to something yousaid before about the data integrity
and the availability of high qualitydata is of paramount concern.
Mm hmm.
I think everybody agrees, right?
Everybody wants good data and more of it.
There has never been an issuewith that, but what is good
data is very hard to know.
just like when we go backto the analogy of art.
You have to see a lot of art tohave, you know, an educated opinion.
Yeah.
to recognize art what it is.
Yeah, but you see more data fromdifferent companies, with different,
I guess, histories and contingenciesand, all that played into it.
I would love to hear your take.
What are the truths about good data ormaybe also the unintuitive lessons, like
what would empower someone to, whetherthey work with an external consultancy,
with AI, or even with, somethinginternally, what do they need to have data
wise and quality wise to be set up well?
Well, so I think the first thing isyou should always have the attitude or
approach of trust but verify, right?
and that's why we have analysts focused onthe QA ing of this information and working
with the client every day to let themtell us if they think this is good data.
Is it from a credible source?
Whether it's your internal data,it's public data, or it's data that's
being, delivered through subscriptionsto certain data sources or even
from our own internal library wherewe're training, on a regular basis
and doing industry level things.
So you've got to start with trust,but verify, Don't just take it for
what it is on its face until you'velearned how to trust it over time.
you do need to have a mechanismin place for, ensuring
quality.
When you say trust, but verify,is that for you a guided process
through a product or a process?
Or how would that pan out?
What does that mean for someone whosays, Oh, we have one petabyte of data.
We're in a perfect position.
I've verified it.
It's one petabyte of data.
We're good.
well, I think that's where you needto have a human in the loop, right?
There will still need to be ahuman in the loop or a set of
humans, the artists if you will.
that have the insights, the knowledge,the lived experiences, the subject matter
expertise, the domain information, andall of the other nuances around the world
in which they operate and move through.
they need to be, acutely aware, of,those outputs and make sure that
they're coming from the right place.
and that it's good information.
And one of the beauties of whatAI can do for you is it can pull
from a lot of different sources sothat you have basis of comparison.
and you can benchmark against work youmay have already done done manually
. for example, and we see that a lot.
as far as, good data structures, look,the end of the day, it's, if you can
crawl it, we can source it, and we cansynthesize and summarize it and give you
an opportunity to digest and then makegood actionable decisions based on that.
that's where we are today.
The future looks like visualrepresentations of that.
A lot of people are asking for, okay,now how do I take all this text based
summaries and insights and createvisual representations that tell
the right compelling story So thatI can get my organization to buy in.
and that's coming, but we're stillworking on the datasets and we'll
always be working on those datasetsbecause that market is changing rapidly,
whether it's Anthropic, of course,or, perplexity pro or open AI, or.
Co pilot, you name it, right?
so you always have to have a mind on that.
and then the results and the sources ofthose results are absolutely critical,
which is why it's critical to make surethat you have, sort of a data supply
chain in place that allows you to trackand trace source of that information and
determine whether or not it's, credible.
That was really impactful as a statementyou said in a way you were in the business
of killing consultants or the businessmodel of management consultancies.
Now, it happens to be the case thatAccenture had 600 millions in booking
for a generative AI related I thinkit's their biggest single line item
for them, and they're on track to makemore than maybe 2 billion this year.
So where's the discrepancy, right?
It seems to be like the hottestand the most flourishing part for
consultancies, gen AI, but yet,you just said this will be killed.
So how do we make sense of that?
Well, I think what you're just seeingis the transition and they're smart and
everybody pivoting towards AI strategyand implementation work because that's
what companies are doing right now.
almost every client we work withhas some sort of quote unquote AI
working group that they put togetherto try and figure out where and how
we should be implementing this stuff.
And so, for the near term, whateverthat looks like, five years, 10
years, that work will be there.
but they're also going to be able todo it much more efficiently with a lot
less, people in their organization.
And so they're going to get leanerand and there will be fewer jobs
and there'll be more output andthere will be pricing changes and
adjustments based on those things.
But over time, that's going to wanebecause if you really are using
AI the way you should be using it.
It should enabling you to dothe work of those consultants.
overwhelming majority of them.
As I said, it's not goingto be 100 percent gone.
But you will have smart, trustedsources leading these things for you.
And C suites will always demand that.
It's their nature.
They want that certainty.
And they want that human experiencedriving this stuff for them.
So, you're just going to doit with a lot less people.
80 percent of it'sgoing to go away, right?
20 percent will remain and work inthat capacity, but they'll be doing it
powered by AI and helping enable their
clients to be powered by AI.
that sounds very plausible, right?
if that's true, what will bethe strategy for companies out
there in what should they invest?
in terms of what can be thedifferentiating competitive
advantage of a company.
And also, since a lot of people willbe in one way or the other affected
by it and turning into artists orhaving to change the core of the work.
How do you see that play out?
what are the right things tosupport or do, especially since we
had other guests on this podcastpointing out that the change.
management process will be ofthe most decisive factors in
turning this into value not.
So your take
on it,
I think from a talent perspective, ifyou want to be relevant in the age of AI,
you, need to become more of a generalist,
and you need to understand how towork from the problem backwards
to leverage things like that.
That's why I talked aboutpairing subject matter experts
with, technical experts, right?
The artists and the scientists.
so to upskill and reskill yourworkforce to be more generalist in
nature is going to be super important.
then, you know, where do you start, right?
you start by where can we createbusiness value with applicable use
cases today, regardless of whether ornot you need to build it or it already
exists in the market and you can buy it.
I'm going to do thethings that either create.
top line, revenue, opportunities, growthe top line or, grow the bottom line.
so make money, save money, create andmaintain a competitive advantage and
stay out of prison as I like to say.
Risk and liability.
and there are very applicable use casesright now across the board using existing
tools that are already in the market.
And then putting your thin veneeron top of them to create that value.
and it's surprisingly more.
Surprisingly, I think to me,probably having been through
so many, look, I've been in thetech space since 96, right, 97.
So, I've been through all thesetransitions from the internet to
com from on prem to SAS, to now AI.
and this one to me has been byfar the easiest one to implement.
because many of the tools exist, tobe able to solve meaningful problems
on a use case level very quickly.
So that's what I'm seeing.
Or scale ups.
Do you see that, this appliesnow equally to smaller companies
and bigger corporations,
It does.
look, at the end of the day, all thisstuff it's not B2C or product or
market or industry or sector specific.
It's what I call H2H.
It's human to human, right?
until bots are buying on your behalffrom other bots who are selling to
them, that's not going to change.
we still want that human experience.
we still want those relationships.
We still need that trust andcertainty, and reliability.
And that's only going to comethrough human experience.
So this whole thing is H2H.
It doesn't matter if it's forinternal use or external use.
And I don't see that changing.
That's a beautiful way of puttingit, but not everything is easy.
So we always like to ask one questionof all the processes you see out
there and the impediments, like if youcould transform or change one of those
processes, which one would it be why?
well, so I'm a big fan of AI for,personal and professional development.
in fact, I've invested on theboard of a company called,
Primer AI, which is an AI.
Lifelong Learning Companion forPersonal Professional Development.
It has a wide variety of applications.
Obviously, it's K 12,universities, and individually.
what it does to help put aperson on the right path to
learn and grow is remarkable.
and I'm very excited about that future.
Is it going to replace teachers?
No.
Is it going to helpteachers be better teachers?
Yes.
Is it going to helpstudents be better students?
Yes.
I'm a big fan of the co pilotapproach towards the use of AI.
Everybody kind of jumps to theend and thinks about full agency
and gets scared by the media andthe hype and everything else.
We're nowhere near that.
There are certain elements that you canprovide full agency to, but it's very low
risk, high volume, task oriented things.
but the things that you and I cantake along with us, just like we
take along our mobile phone, thathelp make our lives better, you know,
that's what I'm most excited about
Well, that's a very upliftingand beautiful message.
And I would say also theperfect message to end on.
Thank you so much, Sean, forbeing part of the podcast today.
Good luck uh, with usheringin Gen AI for all aspects of
corporate, not just innovation.
And with that, thanks for listening in toanother episode of Process Transformers.
Thank you, Lucas.
I only wish the audience could be apart of all of our other conversations
that are lengthy and varied, andI always enjoy talking with you,
so thanks again for having me.
And with that, thanks somuch for listening to another
episode of Process Transformers.
This podcast is brought toyou by the dedicated efforts
and hard work of our team.
So our heartfelt thank you goesout to Beyza Kartal, Dharmik
Rathod, and Reagan Nyandoro.
If you have any questions or comments,email us at processtransformers at sap.
com.
And until next time, foranother transformative
