Are two ex-McKinsey consultants out to kill McKinsey with AI?

0:00Mike Butcher:Hello, welcome to Path Founders with me, Mike Butcher. We're covering both the code, the capital and the consequences. So-called corporate reputation, literally the reputation of a company, now accounts for roughly 30 % of the market value in large public companies, according to some estimates. And that equates to trillions of dollars in value. And yet the systems used to monitor and protect that value are madly fragmented. You've got markets, valuing companies, you've got media, you've got operations, supply chains, everything. So how on earth do you synthesize all of that information? Well, companies like McKinsey are often brought in to assess this kind of thing.

0:44Mike Butcher:Funny that because we've now got two ex-McKinsey consultants applying AI to this very problem. Omniscient is a Paris-based startup building what it calls a decision intelligence platform for boards and senior executives. And I'm joined now by co-founder and CEO, Anno Destian. Thanks very much for joining us on PathFounders. Hey, Mike. Great to be here. Thanks for having me. Well, what a good idea. Yeah, you know, and also we've often seen analysts and people coming out of consultancies and taking up tech and doing startups. But of course, in the era of AI, this potentially has a lot more power, doesn't it?

1:28Mike Butcher:So but tell us what you're actually doing in terms of this problem that used to be called machine learning. Now we're using AI to solve these problems. But what exactly is going to be different about what you're doing? The main difference revolves around the generic revolution really and there was also the call that got us to get started with my co-founder Mehdi. We both coming from McKinsey had seen that common challenge coming back ever and ever and ever again during client engagements and there had been companies trying to create what they call digital boardrooms and they were quite traditional solutions giving a pause on you know live operation metrics what's going on with markets but they really lacked any type of enabler to start contextualizing the data and that's when you know like fast forward transform models happened and it changed the game because for the first time we had a way to build up ground up from like the foundational data layer down to the boardroom really well that's ironic you say that because i'm actually on the road and i'm borrowing someone's boardroom to record this podcast but the the um the actual uh the news story here that part of the reason we're interviewing you now is that you've just raised four four point one million dollars uh from the likes of c camp and others um what do you think attracted uh them to you the broader vision really and that's why you mentioned corporate reputation in the first place And that's our starting point, because what we wanted to do is really cover 360 view, what you'd find in the next executive committee.

3:08So, you know, tackling every decision making vertical and corporate reputation is one of them. Supply chain is one of them. Finance is one of them. And SITCAM really embraced the potential of a broader vision. And so that as the springboard to then enable the value acceleration once we start aggregating the next wave of use cases. And not many funds realized that vision. And we had also a very good connection on a personal level. I think they've got a fantastic team at Seatcamp. So very snappy, got intellectually, emotionally, strategically a very good read on the team.

3:44Mike Butcher:Right. Yes. And yeah, Seatcamp's reputation precedes it as an investor. But what's going to be the end game and the end product here? We've all seen dashboards produced about companies, whether they be large ones or small ones, business intelligence, etc. What ends up being the product? Executive ready insights. You know, that's one critical point of stressing is that we're not another dashboard. We're not another social monitoring platform. These solutions hinged on fairly traditional superficial dashboards. There were exceptional solutions based on the technologies available back at the time.

4:25But what we do now is we get the insight, we extract the insight out of the data sets. So instead of having just basic quantitative metrics and potentially world clouds, what you're getting is very crisp insights, action driven insights, a set of recommendations, everything broken down from the very new five minute condensed executive read to a more granular approach. promotional layer. And that's the promise of generative AI as well, is that the genetic systems are going to allow that flexibility to contextualize the data, essentially mimic the work that was being done manually by the analyst.

5:03And with corporate reputation, what that meant was sifting through thousands, if not millions of articles, social media posts, trying to understand what's happening on the ground, and then extracting the data. And then you multiply it with the number of brands, number of sources, the number of geographies you're covering, and you end up with a very complicated, like very heavy back and forth. And that means time lags, a lot of resources being spent on very large internal teams or external communication agencies, and then a lot of frustration as a result because you're always late.

5:39Mike Butcher:Sure, but we all know that AI also can hallucinate and confabulate. So how are you going to guarantee the accuracy of what you put out at the other end? I mean, is it a matter of just sort of hooking the corporate systems up to your platform? And how do you guarantee accuracy? Yeah, it's a critical point you're bringing on the table. And, you know, quite naively when we got started, we assumed that Gen.I. would have solve it all, it was like for everything into an LLM and got perfect pre-recommendations, analysis. And what we realized is it really depends on the combination of deterministic layers, where we use traditional machine learning for data engineering, entity extraction, sentiment analysis, a number of traditional natural language processing activities before then harnessing the power of GNI.

6:36And then when we start unleashing the power of agentic systems, we also need to ground them into the context of the organization. Of course, as you said, it's brought to hallucination for many different use cases. They start being shaky. So you need to start having a lot of safeguards to put in place, guardrails, whether it's in the agentic systems themselves with LLM as a judge, there are a number of mechanisms that we can start implementing. but it's also about weaving in that process in the user experience and knowing when to call for human validation. And that's why right now, you know, we cannot automate the process end-to-end entirely, but we're going to make the work and the vetting just 100 times more efficient.

7:23And that's the promise. You're getting a much better service, much quicker, and the human analyst or, you know, the CXO is just focusing on the very high value adding piece of the puzzle.

7:34Mike Butcher:Okay, but it's also a crowded space. Surely there's not just you having this idea. How do you feel you're going to be able to differentiate yourself against other AI startups trying to tackle similar problems? Different starting points. I think we've seen a number of incumbents getting started from social monitoring, ESG supply chain. We got started with corporate reputation because we're talking about fast data sets. The question is going to be how quickly can you start capturing value, combining the different modules that we're going to include. And corporate vision, again, is one of them. We started working on additional modules around geopolitics, narrative penetration, which is a more marketing angle.

8:17And the more modules start integrating, the deeper entrenched you become within the organisation, then the strongest is the more going to be. and what we want to do is double down on the experience. We have the benefit of being Gen AI native when a lot of incumbents would have to go through a lot of tech debt. So it's quite hard for them to repurpose the tech stack. And we are quite cautious in trying to keep a close eye on what the other newcomers are proposing. But realistically, we are going to see a lot of them emerge for the next couple of months and years. but we haven't seen anybody really just sealing the deal so far.

9:00So there's a bit of a blue ocean, at least a window that's going to close at some point, but we want to exploit the best of our ability, of course.

9:09Mike Butcher:Sure. I'm sure there's a few other people who've had similar ideas. Do you think you're going to be selling to companies themselves or also you'll end up selling to other analyst companies? Another brilliant question. A bit of both and how could market reflect that? We're pushing on both fronts. Historically, we've been selectively working with companies directly, so our end clients. And now we've started negotiating with communication agencies, services firms. McKinsey would be one of them, of course, but they have the benefit of using that as conveyor belts. most of them are not tech firms. So it's an interest to start collaborating with the right platforms to be able to get the best out of their value proposition, which is very senior advisory and very creative right now activities that do escape some of the automated solutions.

10:07Well, we did actually see earlier this year that an AI company faculty was acquired for a billion

10:15Mike Butcher:by Accenture. So that seems like a natural exit opportunity for companies like yours. I presume obviously it's early days, but I'm sure that you might be thinking about that at some point in time. The short answer is we're not really thinking about it because the value at stake is just huge. And I think the question is, do we have the right platform to go deeper, to go wider? As long as the answer is yes to that question, we'll keep going. There might be a time when we realized that some players are better positioned and we'd be running into a wall. And these would be the moment we start considering acquisitions, but very much aligned right now with capturing as much of that border vision.

10:57We have a lot of clarity, a lot of trust in the direction the market is going to take. The question is how the power of four is going to play. And if the power curve is in our benefit, then we keep pushing as far and as long as we can.

11:13Mike Butcher:Well, very interesting to see the emergence of this model. I'm sure this won't be the last we touch on this subject or hear from you. But for now, let's leave it there. Thanks for joining Path Founders, CEO and co-founder of Omniscience, Arnaud Destin, and see you again. Cheers. Thanks, Mike. See you well soon.