February 27, 2024 • 33 mins
Welcome to this stimulating episode of AI Unleashed, featuring our host Gianni Samwer conversing with Jannik Seger, the Head of AI Operating Model and Strategy at Applied AI. Gain inside knowledge of Jannik's career progression, his role as a dynamic AI leader, and his forward-thinking strategies involved in cementing European industries' place in the global AI market.
This conversation delves deep into Applied AI, a company championing the rise of European industries in AI's competitive landscape. Jannik openly discusses the company's commitment to open exchange and elaborates on how it empowers other companies with modern AI business models and methods.
The episode broadens its horizons by discussing the challenges posed by big tech giants and the growth of AI start-ups. Jannik delivers profound insights about potential shifts in AI jobs and the competitive advantages AI presents. He also highlights his endeavors in increasing the number of use cases transitioning to production.
Finally, the discussion highlights Jannik's foresight on Europe’s position in the global AI scene, its hurdles, and distinctive opportunities. A detailed exchange on strategic partnerships and how Europe could become a powerhouse in AI-driven healthcare solutions finds mention, wrapping up the conversation on the note of worldwide industry reconstruction driven by the transformative power of AI.
Dive deep into the revolutionary potential of AI, effective strategies for AI initiation in businesses, and pivotal considerations regarding the auditioning of AI systems for regulatory and ethical compliance. Historical parallels between the industrial revolution and rising AI underline the importance of continuous learning and strategic decision making amidst evolving AI technologies.
We also touch upon potential concerns about AI replacing human jobs, offering a perspective on how AI can enhance human capabilities and efficiency. Concluding the dialogue, we share critical challenges in the emerging AI industry, including regulation, data quality, and talent availability and offer advice for professionals considering integrating AI to amplify business efficiency, emphasizing the need for setting clear objectives, managing expectations, and building partnerships.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:02):
You are listening to AI Unleashed with your host, Johnny Samur.
Tune in to hear from experts in the field and influential leaders as they share
their expertise on the potential opportunities and challenges of artificial
intelligence and what they believe the future might hold.
Make sure to subscribe wherever you listen to your podcasts.
(00:28):
Hello and welcome back to AI Unleashed. Today, I'm honored to talk to the head
of AI operating model and strategy at Applied AI, a dynamic leader driving AI
innovation and strategy in the European industry.
With a background as an associate at BCG, he brings a wealth of experience in
shaping AI strategies to propel businesses forward. At Applied AI,
(00:50):
my guest is at the forefront of ensuring European industry competitiveness in the global AI market.
With a collaborative approach and a commitment to open exchange,
Applied AI empowers companies to harness the latest AI methods and business
models for scalable and value-creating AI applications.
It is my pleasure in welcoming Yannick Seeger to the podcast.
(01:11):
Hello, Yannick. Thank you for coming out on the podcast. It's a pleasure having you.
I wanted to start off with maybe you just introducing yourself,
what you did before and how you got into the AI field and maybe like sharing
a bit about your journey in the field of AI from your time at like BCG to your
current role as like the head of AI operating model at like Applied AI.
Sorry so hi everyone i'm janek ziga i'm the
(01:33):
head of ai operating model and strategy at appliedai
and we are europe's largest initiative for
the application of trustworthy ai technologies and corporations and as well
thanks a lot for having me on the podcast my journey into ai was actually i
got so the first concert with the i was as an intern at the volkswagen group
(01:57):
center of excellence for autonomous driving.
I was then working in the strategy department, then continue my journey at the
Future Society, which is a think tank that is based at the Robert Kennedy School of Government.
I was interning with them as a research associate and established a framework
that helps intergovernmental organizations.
(02:19):
So think about organizations like the UN,
and the world bank and and so on to leverage ai
in their daily work and then basically joined
dcg as a management consultant after previously interning with them um had a
great time there and advised yeah leading companies across multiple industries
but then yeah decided to follow my calling again for technology and ai and joined
(02:45):
acquire.ai just because i was a believer in the the concept,
the people behind it, and also the unique positioning of how the Ionet joined.
Roughly three years ago, but then as an AI strategist, yeah,
and now we both were senior strategists, was team lead for AI strategy transformation,
now focusing on operating model topics as head of AI operating model within
(03:09):
the strategy and transformation team.
Yeah, that's really interesting. You talked about applied AI and obviously the
work you seek to accomplish there.
What are the primary responsibilities and challenges you face in your role right now.
So in general, we, or I and my team, we work across the full,
(03:29):
what do we call, AI strategy spectrum.
So in general, we focus on ambition development.
That means what's the future competitive advantage that companies can achieve through AI.
Based on that, I'm deriving what are the fields of action. So where should you
play and how do we get commitment, end-of-the-world commitment from senior executives?
Then going down to the use case, ideation, assessment and prioritization.
(03:54):
And lastly, we focus a lot in our work on the enabling factor.
So what does a company need to bring AI into practice?
And that is around organization expertise, but of course also culture, data and technology.
And lastly, the ecosystem you're in. And basically, I would say the main challenge
that we see, and this is also why Applied AI was founded back then,
(04:17):
that it's still quite easy to get to POCs in AI,
but it's super hard really to build AI applications that scale across multiple
parts of the company or multiple jurisdictions.
And this is still our main challenge and the work we're doing on this today.
Do you think that in that case, then the big tech companies like Google and OpenAI,
(04:43):
which now have a big share in the AI field, Do you think that will continue
or with the amount of startups coming up there will be at least like a shift
towards like multiple firms running the space rather than like one firm or like three?
It's a good question. I think in general, what we see is a shift to a more multipolar
(05:07):
large language model world.
I mean, now that you're mentioning OpenAI and the likes, so I think we will
see a switch from having a couple of dominant players right now to,
and I mean, these companies include like Google,
Netflix, OpenAI and the likes, to companies.
And probably that's also what you're referring to with startups to having companies
(05:28):
that offer smaller, more specialized LLMs.
I mean, if we just look into the LLM space, and I think in general,
Gen-AI has been a trend of 2023 and will continue to be the trend of 2024.
But I think in general, one shouldn't just limit the field of AI to Gen-AI or
(05:49):
large language models. What we see, especially in industrial environments,
is still a lot of classical machine learning.
For example, if you look at production settings and so on, and there I think
it still is valuable for companies to build up own skills and not just rely
on large tech players and basically just having an API to them,
(06:09):
but rather building up own models, deploying them.
Yeah, I totally agree. You mentioned how in some areas you can get a valid solution
with just using machine learning.
And I think that's also where a big strength of AI lies in that you can more
or less solve relatively big problems, either to your industry or generally.
(06:32):
The earth faces with relatively or with
not that much work right i i created like
own machine learning models and what what's the main takeaway
was like how easy it is to get like a good
accuracy model going which can then solve like
or make your make your day efficient and
yeah just just solve problems you mentioned
(06:53):
like obviously in in relation to applied
ai obviously the industry in europe how do
you or how would how does your team contribute to like ensure the european industry
remains like competitive in like the global ai market because obviously right
now the the main or like the main contributors to the ai boom were like the
(07:13):
ones obviously in america absolutely.
So in general i think it helps first to see we
as a planet i were split in a for-profit entity
and and also a non-profit entity and within a
for-profit entity we basically offer a
range of solutions and services that is i mean
in our strategy and transformation team as i said ambition development
(07:36):
ideating use cases and then ideally you have a hand over to our what we call
the individual solutions team where we really develop ai products from initial
scoping over our POC development up until a product that is ready to be integrated
into the standard workflow.
So this is one thing. We also support companies through trainings for our capabilities
(08:00):
team and through larger structured programs that basically combine these elements
that I just mentioned into one large transformation program. So this is one thing.
The second pillar is we have a lot of different tools that are out of public
right now or in development that is,
for example, a tool to assess AI maturity that we've rolled out in Germany,
(08:21):
but also across Europe through multiple local players in European countries.
We have developed a risk classifier that allows you to simply assess the risk of your use cases.
With regards to the upcoming uai action then lastly we
also have a tool that focuses on use case management and there
the goal is to increase the number of use cases that
(08:43):
make it to production so this is one pillar the second large pillow
is that we are a partnership of companies
so companies can become members and these companies include large
players like siemens bmw mtu dot
telecom antelites but also large german family-owned
companies or middlestown players and they engage in
(09:04):
exchange with each other and lastly as i mentioned the
non-for-profit we really have
a broad offering of educational courses that we
offer for free and we also publish regularly the
european startup landscape and also have a
lot of information and training material for example on ai
regulation and also our transfer
(09:26):
of trainings for example and other engineering resources so this
is how we really contribute both to companies
um but also to the wide republic how would
you say that in the near future or in the in the future
that europe will have like a big share of the
ai or like will have a big share of ai companies that
are quite successful in relation to obviously america and
(09:48):
obviously china as well so because i mean
right now the all the big tech companies are more or less focused
in america or in in america and.
China and how do you think this will change with like the
ai market and whether europe will let's say have have a big boom and have a
have a good percentage of market share that's an excellent question i think
(10:11):
to be quite honest like the large you know cloud providers and also open ai I mean,
they are quite far ahead.
I mean, there is no denial about that.
However, I think we have a lot of promising startup activity in Europe,
if you look at Aleph Alpha, for example.
But I think in general, I mean, I see a couple of challenges that probably hinder Europe.
(10:35):
And I mean, one is the relatively stringent regulatory environment that we have
in Europe, particularly when it comes to GDPR, that poses a challenge to companies.
Companies, we also have, I would say, a quite severe talent shortage. So there is a,
noticeable gap in AI skilled professionals. Lastly, but I think that's something
(10:55):
that is really changing rapidly is the public perception and trust towards AI technology.
I think in Europe was less than it probably was in the US and China,
but again, I would say that is changing quite rapidly and still.
I mean, something that applies to technology in general in Europe is that Europe
is just a quite fragmented market with a diverse linguist, but also cultural landscape.
(11:17):
However, However, I think one still shouldn't underestimate the opportunities that we have.
And I think in general, we are really positioned as an innovation hub,
as we still have great research institutions.
And I think I would say the other side of the coin of the regulatory landscape
that I mentioned is, of course, that Europe has a focus on ethical and sustainable AI.
(11:39):
And this can really be a significant differentiator in global markets. markets.
Still, I think if we look into the industries, I mean, we still have a huge
opportunity in our traditional industry, the manufacturing base to improve efficiency there.
And also I recently had a discussion on basically advancements in healthcare
(12:00):
because Europe still has quite a role, especially in Germany healthcare system,
which is also we could focus more on this this industry and also position ourselves
as leader in AI-driven healthcare solutions.
And lastly, I think something that functions quite uniquely in Europe is partnerships
(12:22):
that include both the private and the public sector.
For example, one example being the Innovation Party AI that is currently built
up in Heilbronn by the state of Baden-Württemberg and also the Dieter Schwab system.
And they're also one of our shareholders, for example. But I think this is really
a great example of all the private partnerships that work and that drive innovation.
(12:46):
Yeah, one thing you mentioned was obviously that we, as Europe, has a talent shortage.
Why do you think that is? Do you think that, obviously, I feel like a lot of
people move from Europe to America because they think, obviously,
it's like a better ecosystem for them to found something?
Do you think that's the main reason or is it maybe like that the wages are higher
(13:11):
for like the tech companies?
I think it's both factors. I mean, especially if you look at,
I would say, the super high qualified.
And also if you look at startup partners, on the other hand,
I mean, one also has to say, for example, if you just look at the investments that large tech players,
around, be it Apple, be it Google and the likes, I think the overall talent
(13:36):
shortage is more there's so many industrial companies, especially if we look
in in Germany, you know, the classical hidden champions, the classical metal
chance companies and so on.
I think most of them they will really have a hard time of in hiring qualified talents.
And I think you can counteract that to a certain extent, you know,
(13:57):
with recent gradients. And I think that is a great initiative.
However, what we're trying to look at,
and we also have a dedicated program at a point in life where that is
really equipping these graduates with project experience
throughout their studies already so
they so that they're able to yeah implement
cases i would say right away after they graduate so i think in general the talent
(14:20):
shortage is more in the broad industrial base and less at the i would say the
leading players because they of course also you know invest in europe need that
in in Germany, being able to understand what care is in London for us.
Yeah, totally. Obviously we spoke a lot about like the future and how the industry will evolve.
(14:40):
Where do you think, or what are the current key trends in AI and like,
how do these trends shape the strategies of companies across the world and maybe,
obviously one strategy will maybe be applicable to like China,
but obviously not America or Europe.
Yes, so I would say probably two answers. I think one regarding the trends.
(15:04):
As we discussed previously, I think, of course, Gen-AI and LLMs,
they have been trends 2022, 2023, growing with LLMs, of course,
but I think they were really bringing them into application.
That's what I basically see as the main trend for 2024.
So translating them into productive productive applications,
(15:25):
not just playing around with them, so to speak.
I think secondly, from a technical standpoint, I would say we see open source
gaining traction against proprietary models.
And also, if we think about at
least what we experience in discussion is a shift in AI job market and MS.
(15:47):
We have a gradual shift from viewing AI as a job market killer into rather something
that is focused on new job creation or new opportunities.
So I think that is one thing. And then basically, as you were asking for strategies,
I think especially for German companies, what still holds true is that you have
to think about AI really as a holistic transformation.
(16:10):
And there we always, we've just found that our strategy house provides great
guidance to companies because it allows you to set the right direction for the ambition,
to ideate the right use cases, but also to think about the multitude of enabling
factors that I mentioned.
And of course, then all of that, you know, should lead to a successful execution.
(16:33):
And it really should lead to that you can cover all the necessary expertise
or skills required across the full machine learning lifecycle.
So this is one framework that I think is helpful in terms of strategic decision making.
And secondly, I think it also helps in, you know, structuring your initiatives on a time dimension,
we have the framework we call the AI
(16:55):
journey that we differentiate for different maturity
levels on and um i think it
really helps you know especially if you're just
starting off with ai and i think with all the hype around opening ai and so
on we shouldn't forget that there is a large majority of german companies especially
if you factor in the smaller companies i would say at least two-thirds but probably
(17:19):
rather three-quarters of companies they're really just starting off their initiatives initiatives.
And there I think it really helps in you know, structuring your activities across
the strategy house and also thinking what are the dimensions that I need to
focus on first, and which dimensions probably do a better start in a year or two years time.
(17:39):
Yeah totally i think you mentioned how a lot of people are
obviously scared that ai will replace our jobs and i think one thing that's
really important to remember is obviously during the industrial revolution obviously
like a lot of the population was working on like farms and just doing light
like hard physical work and obviously with the rise of machines then,
(18:00):
that that work got replaced but obviously they didn't lose their jobs rather
they were or using the machines to help them.
And I think that's going to be the same with AI.
Rather, obviously, maybe it will replace some jobs. Or like some jobs will definitely
be different. I think there will be more jobs.
But I think definitely right now, maybe a person who writes emails will definitely,
(18:23):
the job won't be the same.
Rather, the person will have to check the emails again or maybe write the prompts
for the emails or something like that.
I think that's one really important thing to remember.
No, absolutely. I mean, there is this famous saying, I think it is definitely
right that, you know, it's not AI that will replace humans or human jobs,
but it's rather a human that knows how to leverage AI just because you have
(18:46):
massive efficiency gains.
Yeah. And also, I mean, one thing I think that is always worth mentioning,
and I remember also Sam Altman said this when he was visiting Munich as part
of the OpenAI World Tour, where experts really have a hard time in predicting.
Which jobs will become obsolete or will be partly replaced by AI.
(19:08):
I mean, if you look into studies that are probably three to five years old,
it's quite amazing to see how really smart people with really dedicated research,
what their predictions were, and then also the kind of jobs that we see at least
partially automated right now,
largely we're into the ones that have been predicted to be automated so it is
(19:29):
hard to foresee that even for experts but i think you're absolutely right it's
you should always perceive it as an augmenting technology and that definitely
will lead to efficiency gains and i think the jobs will be replaced it will always be,
or in most cases a human interacting with ai systems or overseeing these ai
(19:49):
systems Yeah, I totally agree.
Obviously, now with the rise of AI the most and it making the work obviously way more efficient,
where do you think or how do you think government companies can navigate the
evolving world around AI to stay ahead of the market? Because obviously.
Like companies that use AI will definitely be more efficient and therefore like
(20:14):
get more work done than obviously the companies that won't get work done.
Yeah, so I think what I mean, our work with the companies who we advise and
also the partners that are part of the initiative,
I think really important if you start out with AI, but also,
you know, just to regularly check the straight line, clear objectives,
(20:35):
like what do I want to achieve with AI?
AI and how does AI benefit the overall mission that I am on as a company?
I think secondly, you really have to ensure adequate data quality,
both for mobile training, but then of course, also if you run these models.
I think that is definitely what I would see as a second priority.
(20:58):
Thirdly, I mean, we've talked a lot about the talent shortage is really about
how do we develop the right talent? That could both be through internal qualification
or upskilling or external hiring.
And then I think also, if you want to apply AI successfully,
you need to foster a culture that also allows that.
So a culture that values continuous learning, but also experimentation,
(21:21):
that is able to adapt to new AI technologies and methods. efforts.
I think also something that has proven really successful, irrespective of the
industry or the company size, just to start small and scale gradually.
And most companies, of course, they initiate their AI integration with small
(21:41):
pilot projects, test the negative
impact, and then basically apply these learnings to larger projects.
I think it's also super important not to over-promise at the beginning.
And then lastly, I think this has been true for players that we work with.
Almost all of them they really engage in partnerships with ai vendors with academic.
(22:02):
Academia or academic institutions and also
with industry groups such as applied because i'm really
convinced that these collaborations provide insight
or yeah they basically allow you you know not
to repeat the mistakes that probably peers in other industries
have already made and also give you of course a
perspective on the trends and
(22:24):
best practices and especially i think regarding out in a transit it
is hard for a single company to
stay ahead of the curve just because the field
evolves at a speed that is probably you
know unheard of in in all the technology development
yeah definitely i mean we we obviously talked a lot about like implementing
(22:44):
ai and how you can use it to your advantage where do you currently see like
the biggest challenges and opportunity And like challenges for like AI adoption
and development in particular, in Europe particularly.
And especially one thing that I could think of right now is I feel like right
now we're focusing on too much of like regulating AI rather because I feel like
(23:07):
we're kind of treating it as if it's already like conscious and has this AGI
part to it where obviously we're not there yet.
So obviously they're trying to regulate it, maybe regulate the data you're using
or regulate how big your models can be or how accurate your models can be.
And obviously that is like decreasing the chance of obviously maximizing your product.
(23:33):
Sure. I mean, besides, you know, the core factors that I was mentioning previously,
or let me just repeat, I think overall for companies or layers of companies, it is data,
data, data, and it's core talent as well.
But then, you know, if we look at some more nuanced factors,
I think one is that you really need, you know, the buy-in and the actual real
(24:01):
tangible buy-in of executives.
Executives, something that we have recognized in the past is that a major challenge
is that people underestimate how much effort it is to implement AI products,
and especially in corporate environments.
Can be that there is just departments that are unwilling to devote,
(24:21):
for example, the main experts to AI development, so I think this is one thing.
And then secondly, I think what I also mentioned at the beginning,
at the core challenge that we still see ourselves component with them that it's
just really hard to move from a poc stage to.
To production and to a scaling stage i
(24:43):
think there it also again if we just think
about some executives it also is a lot about you know allowing this cultural
experimentation but also um we have dealt with companies that had you know like
we we scored them with our maturity assessment and they had great talent also
their data and technology and scores were quite high,
(25:04):
but we just had a lack of buy in from top management.
So I think that is again, something that I would mention and then I think, you know.
Lastly, having a smart make-or-buy strategy is also something that poses a challenge.
Sometimes companies are unsure which components of an AI system do I have to
(25:25):
do on my own because it really creates a strategic advantage versus which components
would I probably rather source outside.
And I think overcoming this challenge can be done,
first of all, through focusing on the right use cases, but then
secondly also through employing a smart
make or buy strategy because you definitely do not
have to you know you do not or you cannot develop
(25:47):
everything on your own and i think this is something that especially german
companies might have to overcome so the not invention here syndrome is still
strongly at least in certain industries or within certain players in certain
industries and so i think besides the major factors that we mentioned previously
i think these would be the ones,
because ultimately, what you want to do is, you know, you want to have AI as one solution approach,
(26:14):
where it makes no sense, you don't have to apply AI all the time.
And I think one client of ours, they mentioned, and I think this is a great
way of thinking about AI.
They said that they want to have implied have AI within the mental solution
toolkit of employees and have it used as frequently as excellence today.
(26:35):
And I think this is obviously no it's a long way
of going there but i think by having this mindset you can
basically counteract a lot of challenges or
of uncertainty that people see themselves confronted
with when especially when starting off their efforts
yeah definitely do you
think however that we're going in the right direction with
(26:56):
like ai rules and regulation or do you think obviously it's
it's a bit too early to talk about like regulating data
regulating the models those regulating whatever they're regulating i would say
let's say personally the only answer can be it depends i think it depends a
lot on whom you're asking and you know there is valid arguments both like on
(27:17):
the pro ones and then also the con side yeah i mean something that i.
Recognize is i think the ai act and also the current uncertainty that still
you know surrounds around the AI Act and really poses massive challenges to
companies irrespective of the size.
And something that I recognize is, I think in the end, you know,
(27:40):
it's legislation that one has to comply with.
But interestingly, people tend to, yeah, have too much respect for it.
They just say it's too complicated.
And I think breaking that down and really making sure that, you know,
not just the big companies companies that obviously have great in-house departments
(28:02):
and work with great law firms,
but also how do we make sure that the broad space we think about.
Like smaller industrials, but
also startups and the likes, how they can comply with these challenges.
I think that is a question that we have to ask ourselves, I mean,
just because you have to comply with the regulation.
But also, I think it poses a great opportunity in both for existing certification
(28:27):
bodies, but also for other startups or players that move into that. that.
And I think overall, how the regulation will play out is really just something
that we will see once the rules, you know, by and once the rules are actually enforced, I think,
right now, it is just too early to have a definite answer on how positive that
(28:49):
actually is for Europe or the European industry,
or how much it actually hinders development.
I think there is really well it's chosen common on both sides.
I feel like like I said before that
would that we're trying to regulate AI a bit
too much I feel like obviously a bit needs to be regulated but
let AI develop some more and
(29:10):
then once once we reach this AGI phase then then
we can sure regulate how you how you can use it obviously we don't want any
anyone using it for the wrong reasons I wanted to end off on the note like asking
you maybe like What advice would you offer to professionals and businesses looking
to integrate AI into their strategies and obviously making their companies more efficient?
(29:35):
I think, I mean, looking back at what, you know, or synthesizing what I had
said previously, you know, setting clear objectives.
Where do I want to leverage AI? Also, how do I want to do it?
And also having realistic expectations towards the timeline.
I think AI is neither a silver bullet nor is it something that you would employ within three months.
(29:57):
Then also thinking again, how do we really get the talent boards that we need?
How do we ensure buy-in through basically starting small and then scaling things up?
Which partners do we want to work with?
How do we use like, how do we come up with with make or buy strategies?
(30:18):
And then I think lastly, of course, this can be a differentiating factor once
you have reached a certain maturity.
The question is also how do we enable our units?
Because most companies, they start off with a central center of excellence and
that center of excellence basically forms the nucleus for all the activities in the company.
(30:39):
However, I think what you should really plan on early on is how do we ensure
that also our business units, departments, however you have structured your organization.
How do we make sure that they are actually able to operate and maintain the
AI products in the first place?
And then secondly, also how do they develop them on your own?
(31:01):
So I think how do we distribute AI know-how across the whole more entrant machine
learning I think that is also something that is absolutely crucial because otherwise,
you know, you might see yourself confronted with the problem that the handover
of the solution that the Center of Excellence has built doesn't work as planned
and this can really lead to frustration or also to expectations that have not been met.
(31:27):
So I think that is really important to plan it early on.
And I think lastly, I also want to mention how do we ensure that,
you know, we're not only regulatory compliant but also
which ethical guidelines and standards do
we want to follow and this could be in general
accepted standards on diversity and inclusivity but also how do we as a company
(31:49):
ensure that our ai systems are transparent and explainable wherever that is
technically possible and also how do we ensure and again this probably applies
more to companies that that are more mature,
that have actually deployed multiple AI solutions?
How do we perform regular ethical audits and also impact assessments of our
(32:11):
solutions that are in practice?
I think this is something that more mature companies can but also should focus on.
Yeah, definitely. I feel like that's great advice,
especially because AI right now is relatively new and And just people,
I feel like people who are not really in the tech world or like involved in
(32:32):
the tech world are like figuring out how they can use AI.
And I think that helps a lot. So thank you for coming out on the podcast.
Yannick, this was really great. Thank you. You're welcome, Johnny.
Music.
You are listening to AI Unleashed with your host, Johnny Samur.
Tune in to hear from experts in the field and influential leaders as they share
their expertise on the potential opportunities and challenges of artificial
intelligence and what they believe the future might hold.
Make sure to subscribe wherever you listen to your podcasts.
(00:28):
Hello and welcome back to AI Unleashed. Today, I'm honored to talk to the head
of AI operating model and strategy at Applied AI, a dynamic leader driving AI
innovation and strategy in the European industry.
With a background as an associate at BCG, he brings a wealth of experience in
shaping AI strategies to propel businesses forward. At Applied AI,
(00:50):
my guest is at the forefront of ensuring European industry competitiveness in the global AI market.
With a collaborative approach and a commitment to open exchange,
Applied AI empowers companies to harness the latest AI methods and business
models for scalable and value-creating AI applications.
It is my pleasure in welcoming Yannick Seeger to the podcast.
(01:11):
Hello, Yannick. Thank you for coming out on the podcast. It's a pleasure having you.
I wanted to start off with maybe you just introducing yourself,
what you did before and how you got into the AI field and maybe like sharing
a bit about your journey in the field of AI from your time at like BCG to your
current role as like the head of AI operating model at like Applied AI.
Sorry so hi everyone i'm janek ziga i'm the
(01:33):
head of ai operating model and strategy at appliedai
and we are europe's largest initiative for
the application of trustworthy ai technologies and corporations and as well
thanks a lot for having me on the podcast my journey into ai was actually i
got so the first concert with the i was as an intern at the volkswagen group
(01:57):
center of excellence for autonomous driving.
I was then working in the strategy department, then continue my journey at the
Future Society, which is a think tank that is based at the Robert Kennedy School of Government.
I was interning with them as a research associate and established a framework
that helps intergovernmental organizations.
(02:19):
So think about organizations like the UN,
and the world bank and and so on to leverage ai
in their daily work and then basically joined
dcg as a management consultant after previously interning with them um had a
great time there and advised yeah leading companies across multiple industries
but then yeah decided to follow my calling again for technology and ai and joined
(02:45):
acquire.ai just because i was a believer in the the concept,
the people behind it, and also the unique positioning of how the Ionet joined.
Roughly three years ago, but then as an AI strategist, yeah,
and now we both were senior strategists, was team lead for AI strategy transformation,
now focusing on operating model topics as head of AI operating model within
(03:09):
the strategy and transformation team.
Yeah, that's really interesting. You talked about applied AI and obviously the
work you seek to accomplish there.
What are the primary responsibilities and challenges you face in your role right now.
So in general, we, or I and my team, we work across the full,
(03:29):
what do we call, AI strategy spectrum.
So in general, we focus on ambition development.
That means what's the future competitive advantage that companies can achieve through AI.
Based on that, I'm deriving what are the fields of action. So where should you
play and how do we get commitment, end-of-the-world commitment from senior executives?
Then going down to the use case, ideation, assessment and prioritization.
(03:54):
And lastly, we focus a lot in our work on the enabling factor.
So what does a company need to bring AI into practice?
And that is around organization expertise, but of course also culture, data and technology.
And lastly, the ecosystem you're in. And basically, I would say the main challenge
that we see, and this is also why Applied AI was founded back then,
(04:17):
that it's still quite easy to get to POCs in AI,
but it's super hard really to build AI applications that scale across multiple
parts of the company or multiple jurisdictions.
And this is still our main challenge and the work we're doing on this today.
Do you think that in that case, then the big tech companies like Google and OpenAI,
(04:43):
which now have a big share in the AI field, Do you think that will continue
or with the amount of startups coming up there will be at least like a shift
towards like multiple firms running the space rather than like one firm or like three?
It's a good question. I think in general, what we see is a shift to a more multipolar
(05:07):
large language model world.
I mean, now that you're mentioning OpenAI and the likes, so I think we will
see a switch from having a couple of dominant players right now to,
and I mean, these companies include like Google,
Netflix, OpenAI and the likes, to companies.
And probably that's also what you're referring to with startups to having companies
(05:28):
that offer smaller, more specialized LLMs.
I mean, if we just look into the LLM space, and I think in general,
Gen-AI has been a trend of 2023 and will continue to be the trend of 2024.
But I think in general, one shouldn't just limit the field of AI to Gen-AI or
(05:49):
large language models. What we see, especially in industrial environments,
is still a lot of classical machine learning.
For example, if you look at production settings and so on, and there I think
it still is valuable for companies to build up own skills and not just rely
on large tech players and basically just having an API to them,
(06:09):
but rather building up own models, deploying them.
Yeah, I totally agree. You mentioned how in some areas you can get a valid solution
with just using machine learning.
And I think that's also where a big strength of AI lies in that you can more
or less solve relatively big problems, either to your industry or generally.
(06:32):
The earth faces with relatively or with
not that much work right i i created like
own machine learning models and what what's the main takeaway
was like how easy it is to get like a good
accuracy model going which can then solve like
or make your make your day efficient and
yeah just just solve problems you mentioned
(06:53):
like obviously in in relation to applied
ai obviously the industry in europe how do
you or how would how does your team contribute to like ensure the european industry
remains like competitive in like the global ai market because obviously right
now the the main or like the main contributors to the ai boom were like the
(07:13):
ones obviously in america absolutely.
So in general i think it helps first to see we
as a planet i were split in a for-profit entity
and and also a non-profit entity and within a
for-profit entity we basically offer a
range of solutions and services that is i mean
in our strategy and transformation team as i said ambition development
(07:36):
ideating use cases and then ideally you have a hand over to our what we call
the individual solutions team where we really develop ai products from initial
scoping over our POC development up until a product that is ready to be integrated
into the standard workflow.
So this is one thing. We also support companies through trainings for our capabilities
(08:00):
team and through larger structured programs that basically combine these elements
that I just mentioned into one large transformation program. So this is one thing.
The second pillar is we have a lot of different tools that are out of public
right now or in development that is,
for example, a tool to assess AI maturity that we've rolled out in Germany,
(08:21):
but also across Europe through multiple local players in European countries.
We have developed a risk classifier that allows you to simply assess the risk of your use cases.
With regards to the upcoming uai action then lastly we
also have a tool that focuses on use case management and there
the goal is to increase the number of use cases that
(08:43):
make it to production so this is one pillar the second large pillow
is that we are a partnership of companies
so companies can become members and these companies include large
players like siemens bmw mtu dot
telecom antelites but also large german family-owned
companies or middlestown players and they engage in
(09:04):
exchange with each other and lastly as i mentioned the
non-for-profit we really have
a broad offering of educational courses that we
offer for free and we also publish regularly the
european startup landscape and also have a
lot of information and training material for example on ai
regulation and also our transfer
(09:26):
of trainings for example and other engineering resources so this
is how we really contribute both to companies
um but also to the wide republic how would
you say that in the near future or in the in the future
that europe will have like a big share of the
ai or like will have a big share of ai companies that
are quite successful in relation to obviously america and
(09:48):
obviously china as well so because i mean
right now the all the big tech companies are more or less focused
in america or in in america and.
China and how do you think this will change with like the
ai market and whether europe will let's say have have a big boom and have a
have a good percentage of market share that's an excellent question i think
(10:11):
to be quite honest like the large you know cloud providers and also open ai I mean,
they are quite far ahead.
I mean, there is no denial about that.
However, I think we have a lot of promising startup activity in Europe,
if you look at Aleph Alpha, for example.
But I think in general, I mean, I see a couple of challenges that probably hinder Europe.
(10:35):
And I mean, one is the relatively stringent regulatory environment that we have
in Europe, particularly when it comes to GDPR, that poses a challenge to companies.
Companies, we also have, I would say, a quite severe talent shortage. So there is a,
noticeable gap in AI skilled professionals. Lastly, but I think that's something
(10:55):
that is really changing rapidly is the public perception and trust towards AI technology.
I think in Europe was less than it probably was in the US and China,
but again, I would say that is changing quite rapidly and still.
I mean, something that applies to technology in general in Europe is that Europe
is just a quite fragmented market with a diverse linguist, but also cultural landscape.
(11:17):
However, However, I think one still shouldn't underestimate the opportunities that we have.
And I think in general, we are really positioned as an innovation hub,
as we still have great research institutions.
And I think I would say the other side of the coin of the regulatory landscape
that I mentioned is, of course, that Europe has a focus on ethical and sustainable AI.
(11:39):
And this can really be a significant differentiator in global markets. markets.
Still, I think if we look into the industries, I mean, we still have a huge
opportunity in our traditional industry, the manufacturing base to improve efficiency there.
And also I recently had a discussion on basically advancements in healthcare
(12:00):
because Europe still has quite a role, especially in Germany healthcare system,
which is also we could focus more on this this industry and also position ourselves
as leader in AI-driven healthcare solutions.
And lastly, I think something that functions quite uniquely in Europe is partnerships
(12:22):
that include both the private and the public sector.
For example, one example being the Innovation Party AI that is currently built
up in Heilbronn by the state of Baden-Württemberg and also the Dieter Schwab system.
And they're also one of our shareholders, for example. But I think this is really
a great example of all the private partnerships that work and that drive innovation.
(12:46):
Yeah, one thing you mentioned was obviously that we, as Europe, has a talent shortage.
Why do you think that is? Do you think that, obviously, I feel like a lot of
people move from Europe to America because they think, obviously,
it's like a better ecosystem for them to found something?
Do you think that's the main reason or is it maybe like that the wages are higher
(13:11):
for like the tech companies?
I think it's both factors. I mean, especially if you look at,
I would say, the super high qualified.
And also if you look at startup partners, on the other hand,
I mean, one also has to say, for example, if you just look at the investments that large tech players,
around, be it Apple, be it Google and the likes, I think the overall talent
(13:36):
shortage is more there's so many industrial companies, especially if we look
in in Germany, you know, the classical hidden champions, the classical metal
chance companies and so on.
I think most of them they will really have a hard time of in hiring qualified talents.
And I think you can counteract that to a certain extent, you know,
(13:57):
with recent gradients. And I think that is a great initiative.
However, what we're trying to look at,
and we also have a dedicated program at a point in life where that is
really equipping these graduates with project experience
throughout their studies already so
they so that they're able to yeah implement
cases i would say right away after they graduate so i think in general the talent
(14:20):
shortage is more in the broad industrial base and less at the i would say the
leading players because they of course also you know invest in europe need that
in in Germany, being able to understand what care is in London for us.
Yeah, totally. Obviously we spoke a lot about like the future and how the industry will evolve.
(14:40):
Where do you think, or what are the current key trends in AI and like,
how do these trends shape the strategies of companies across the world and maybe,
obviously one strategy will maybe be applicable to like China,
but obviously not America or Europe.
Yes, so I would say probably two answers. I think one regarding the trends.
(15:04):
As we discussed previously, I think, of course, Gen-AI and LLMs,
they have been trends 2022, 2023, growing with LLMs, of course,
but I think they were really bringing them into application.
That's what I basically see as the main trend for 2024.
So translating them into productive productive applications,
(15:25):
not just playing around with them, so to speak.
I think secondly, from a technical standpoint, I would say we see open source
gaining traction against proprietary models.
And also, if we think about at
least what we experience in discussion is a shift in AI job market and MS.
(15:47):
We have a gradual shift from viewing AI as a job market killer into rather something
that is focused on new job creation or new opportunities.
So I think that is one thing. And then basically, as you were asking for strategies,
I think especially for German companies, what still holds true is that you have
to think about AI really as a holistic transformation.
(16:10):
And there we always, we've just found that our strategy house provides great
guidance to companies because it allows you to set the right direction for the ambition,
to ideate the right use cases, but also to think about the multitude of enabling
factors that I mentioned.
And of course, then all of that, you know, should lead to a successful execution.
(16:33):
And it really should lead to that you can cover all the necessary expertise
or skills required across the full machine learning lifecycle.
So this is one framework that I think is helpful in terms of strategic decision making.
And secondly, I think it also helps in, you know, structuring your initiatives on a time dimension,
we have the framework we call the AI
(16:55):
journey that we differentiate for different maturity
levels on and um i think it
really helps you know especially if you're just
starting off with ai and i think with all the hype around opening ai and so
on we shouldn't forget that there is a large majority of german companies especially
if you factor in the smaller companies i would say at least two-thirds but probably
(17:19):
rather three-quarters of companies they're really just starting off their initiatives initiatives.
And there I think it really helps in you know, structuring your activities across
the strategy house and also thinking what are the dimensions that I need to
focus on first, and which dimensions probably do a better start in a year or two years time.
(17:39):
Yeah totally i think you mentioned how a lot of people are
obviously scared that ai will replace our jobs and i think one thing that's
really important to remember is obviously during the industrial revolution obviously
like a lot of the population was working on like farms and just doing light
like hard physical work and obviously with the rise of machines then,
(18:00):
that that work got replaced but obviously they didn't lose their jobs rather
they were or using the machines to help them.
And I think that's going to be the same with AI.
Rather, obviously, maybe it will replace some jobs. Or like some jobs will definitely
be different. I think there will be more jobs.
But I think definitely right now, maybe a person who writes emails will definitely,
(18:23):
the job won't be the same.
Rather, the person will have to check the emails again or maybe write the prompts
for the emails or something like that.
I think that's one really important thing to remember.
No, absolutely. I mean, there is this famous saying, I think it is definitely
right that, you know, it's not AI that will replace humans or human jobs,
but it's rather a human that knows how to leverage AI just because you have
(18:46):
massive efficiency gains.
Yeah. And also, I mean, one thing I think that is always worth mentioning,
and I remember also Sam Altman said this when he was visiting Munich as part
of the OpenAI World Tour, where experts really have a hard time in predicting.
Which jobs will become obsolete or will be partly replaced by AI.
(19:08):
I mean, if you look into studies that are probably three to five years old,
it's quite amazing to see how really smart people with really dedicated research,
what their predictions were, and then also the kind of jobs that we see at least
partially automated right now,
largely we're into the ones that have been predicted to be automated so it is
(19:29):
hard to foresee that even for experts but i think you're absolutely right it's
you should always perceive it as an augmenting technology and that definitely
will lead to efficiency gains and i think the jobs will be replaced it will always be,
or in most cases a human interacting with ai systems or overseeing these ai
(19:49):
systems Yeah, I totally agree.
Obviously, now with the rise of AI the most and it making the work obviously way more efficient,
where do you think or how do you think government companies can navigate the
evolving world around AI to stay ahead of the market? Because obviously.
Like companies that use AI will definitely be more efficient and therefore like
(20:14):
get more work done than obviously the companies that won't get work done.
Yeah, so I think what I mean, our work with the companies who we advise and
also the partners that are part of the initiative,
I think really important if you start out with AI, but also,
you know, just to regularly check the straight line, clear objectives,
(20:35):
like what do I want to achieve with AI?
AI and how does AI benefit the overall mission that I am on as a company?
I think secondly, you really have to ensure adequate data quality,
both for mobile training, but then of course, also if you run these models.
I think that is definitely what I would see as a second priority.
(20:58):
Thirdly, I mean, we've talked a lot about the talent shortage is really about
how do we develop the right talent? That could both be through internal qualification
or upskilling or external hiring.
And then I think also, if you want to apply AI successfully,
you need to foster a culture that also allows that.
So a culture that values continuous learning, but also experimentation,
(21:21):
that is able to adapt to new AI technologies and methods. efforts.
I think also something that has proven really successful, irrespective of the
industry or the company size, just to start small and scale gradually.
And most companies, of course, they initiate their AI integration with small
(21:41):
pilot projects, test the negative
impact, and then basically apply these learnings to larger projects.
I think it's also super important not to over-promise at the beginning.
And then lastly, I think this has been true for players that we work with.
Almost all of them they really engage in partnerships with ai vendors with academic.
(22:02):
Academia or academic institutions and also
with industry groups such as applied because i'm really
convinced that these collaborations provide insight
or yeah they basically allow you you know not
to repeat the mistakes that probably peers in other industries
have already made and also give you of course a
perspective on the trends and
(22:24):
best practices and especially i think regarding out in a transit it
is hard for a single company to
stay ahead of the curve just because the field
evolves at a speed that is probably you
know unheard of in in all the technology development
yeah definitely i mean we we obviously talked a lot about like implementing
(22:44):
ai and how you can use it to your advantage where do you currently see like
the biggest challenges and opportunity And like challenges for like AI adoption
and development in particular, in Europe particularly.
And especially one thing that I could think of right now is I feel like right
now we're focusing on too much of like regulating AI rather because I feel like
(23:07):
we're kind of treating it as if it's already like conscious and has this AGI
part to it where obviously we're not there yet.
So obviously they're trying to regulate it, maybe regulate the data you're using
or regulate how big your models can be or how accurate your models can be.
And obviously that is like decreasing the chance of obviously maximizing your product.
(23:33):
Sure. I mean, besides, you know, the core factors that I was mentioning previously,
or let me just repeat, I think overall for companies or layers of companies, it is data,
data, data, and it's core talent as well.
But then, you know, if we look at some more nuanced factors,
I think one is that you really need, you know, the buy-in and the actual real
(24:01):
tangible buy-in of executives.
Executives, something that we have recognized in the past is that a major challenge
is that people underestimate how much effort it is to implement AI products,
and especially in corporate environments.
Can be that there is just departments that are unwilling to devote,
(24:21):
for example, the main experts to AI development, so I think this is one thing.
And then secondly, I think what I also mentioned at the beginning,
at the core challenge that we still see ourselves component with them that it's
just really hard to move from a poc stage to.
To production and to a scaling stage i
(24:43):
think there it also again if we just think
about some executives it also is a lot about you know allowing this cultural
experimentation but also um we have dealt with companies that had you know like
we we scored them with our maturity assessment and they had great talent also
their data and technology and scores were quite high,
(25:04):
but we just had a lack of buy in from top management.
So I think that is again, something that I would mention and then I think, you know.
Lastly, having a smart make-or-buy strategy is also something that poses a challenge.
Sometimes companies are unsure which components of an AI system do I have to
(25:25):
do on my own because it really creates a strategic advantage versus which components
would I probably rather source outside.
And I think overcoming this challenge can be done,
first of all, through focusing on the right use cases, but then
secondly also through employing a smart
make or buy strategy because you definitely do not
have to you know you do not or you cannot develop
(25:47):
everything on your own and i think this is something that especially german
companies might have to overcome so the not invention here syndrome is still
strongly at least in certain industries or within certain players in certain
industries and so i think besides the major factors that we mentioned previously
i think these would be the ones,
because ultimately, what you want to do is, you know, you want to have AI as one solution approach,
(26:14):
where it makes no sense, you don't have to apply AI all the time.
And I think one client of ours, they mentioned, and I think this is a great
way of thinking about AI.
They said that they want to have implied have AI within the mental solution
toolkit of employees and have it used as frequently as excellence today.
(26:35):
And I think this is obviously no it's a long way
of going there but i think by having this mindset you can
basically counteract a lot of challenges or
of uncertainty that people see themselves confronted
with when especially when starting off their efforts
yeah definitely do you
think however that we're going in the right direction with
(26:56):
like ai rules and regulation or do you think obviously it's
it's a bit too early to talk about like regulating data
regulating the models those regulating whatever they're regulating i would say
let's say personally the only answer can be it depends i think it depends a
lot on whom you're asking and you know there is valid arguments both like on
(27:17):
the pro ones and then also the con side yeah i mean something that i.
Recognize is i think the ai act and also the current uncertainty that still
you know surrounds around the AI Act and really poses massive challenges to
companies irrespective of the size.
And something that I recognize is, I think in the end, you know,
(27:40):
it's legislation that one has to comply with.
But interestingly, people tend to, yeah, have too much respect for it.
They just say it's too complicated.
And I think breaking that down and really making sure that, you know,
not just the big companies companies that obviously have great in-house departments
(28:02):
and work with great law firms,
but also how do we make sure that the broad space we think about.
Like smaller industrials, but
also startups and the likes, how they can comply with these challenges.
I think that is a question that we have to ask ourselves, I mean,
just because you have to comply with the regulation.
But also, I think it poses a great opportunity in both for existing certification
(28:27):
bodies, but also for other startups or players that move into that. that.
And I think overall, how the regulation will play out is really just something
that we will see once the rules, you know, by and once the rules are actually enforced, I think,
right now, it is just too early to have a definite answer on how positive that
(28:49):
actually is for Europe or the European industry,
or how much it actually hinders development.
I think there is really well it's chosen common on both sides.
I feel like like I said before that
would that we're trying to regulate AI a bit
too much I feel like obviously a bit needs to be regulated but
let AI develop some more and
(29:10):
then once once we reach this AGI phase then then
we can sure regulate how you how you can use it obviously we don't want any
anyone using it for the wrong reasons I wanted to end off on the note like asking
you maybe like What advice would you offer to professionals and businesses looking
to integrate AI into their strategies and obviously making their companies more efficient?
(29:35):
I think, I mean, looking back at what, you know, or synthesizing what I had
said previously, you know, setting clear objectives.
Where do I want to leverage AI? Also, how do I want to do it?
And also having realistic expectations towards the timeline.
I think AI is neither a silver bullet nor is it something that you would employ within three months.
(29:57):
Then also thinking again, how do we really get the talent boards that we need?
How do we ensure buy-in through basically starting small and then scaling things up?
Which partners do we want to work with?
How do we use like, how do we come up with with make or buy strategies?
(30:18):
And then I think lastly, of course, this can be a differentiating factor once
you have reached a certain maturity.
The question is also how do we enable our units?
Because most companies, they start off with a central center of excellence and
that center of excellence basically forms the nucleus for all the activities in the company.
(30:39):
However, I think what you should really plan on early on is how do we ensure
that also our business units, departments, however you have structured your organization.
How do we make sure that they are actually able to operate and maintain the
AI products in the first place?
And then secondly, also how do they develop them on your own?
(31:01):
So I think how do we distribute AI know-how across the whole more entrant machine
learning I think that is also something that is absolutely crucial because otherwise,
you know, you might see yourself confronted with the problem that the handover
of the solution that the Center of Excellence has built doesn't work as planned
and this can really lead to frustration or also to expectations that have not been met.
(31:27):
So I think that is really important to plan it early on.
And I think lastly, I also want to mention how do we ensure that,
you know, we're not only regulatory compliant but also
which ethical guidelines and standards do
we want to follow and this could be in general
accepted standards on diversity and inclusivity but also how do we as a company
(31:49):
ensure that our ai systems are transparent and explainable wherever that is
technically possible and also how do we ensure and again this probably applies
more to companies that that are more mature,
that have actually deployed multiple AI solutions?
How do we perform regular ethical audits and also impact assessments of our
(32:11):
solutions that are in practice?
I think this is something that more mature companies can but also should focus on.
Yeah, definitely. I feel like that's great advice,
especially because AI right now is relatively new and And just people,
I feel like people who are not really in the tech world or like involved in
(32:32):
the tech world are like figuring out how they can use AI.
And I think that helps a lot. So thank you for coming out on the podcast.
Yannick, this was really great. Thank you. You're welcome, Johnny.
Music.
Popular Podcasts
Amy Robach & T.J. Holmes present: Aubrey O’Day, Covering the Diddy Trial
Introducing… Aubrey O’Day Diddy’s former protege, television personality, platinum selling music artist, Danity Kane alum Aubrey O’Day joins veteran journalists Amy Robach and TJ Holmes to provide a unique perspective on the trial that has captivated the attention of the nation. Join them throughout the trial as they discuss, debate, and dissect every detail, every aspect of the proceedings. Aubrey will offer her opinions and expertise, as only she is qualified to do given her first-hand knowledge. From her days on Making the Band, as she emerged as the breakout star, the truth of the situation would be the opposite of the glitz and glamour. Listen throughout every minute of the trial, for this exclusive coverage. Amy Robach and TJ Holmes present Aubrey O’Day, Covering the Diddy Trial, an iHeartRadio podcast.
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
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy And Charlamagne Tha God!