In this insightful podcast episode, hosted by Michael Bernzweig, Larissa Schneider, co-founder of Unframe, explores the current landscape of enterprise AI adoption. She highlights the challenges businesses face in extracting value from AI, the importance of tailored solutions, and the shift towards customer-centric AI platforms. Schneider also shares predictions for the future of AI in enterprises, emphasizing the need for outcome-based pricing models and the decline of DIY solutions. The conversation concludes with a look at how Unframe's approach is helping enterprises achieve significant efficiencies and scale without increasing headcount.
In this insightful podcast episode, hosted by Michael Bernzweig, Larissa Schneider, co-founder of Unframe, explores the current landscape of enterprise AI adoption. She highlights the challenges businesses face in extracting value from AI, the importance of tailored solutions, and the shift towards customer-centric AI platforms. Schneider also shares predictions for the future of AI in enterprises, emphasizing the need for outcome-based pricing models and the decline of DIY solutions. The conversation concludes with a look at how Unframe's approach is helping enterprises achieve significant efficiencies and scale without increasing headcount.
Takeaways
Sound bites
Chapters
Michael Bernzweig (00:06.04)
I hope everyone enjoyed that last segment. And coming up next, have Larissa Schneider. She is the co-founder of Unframe, a managed AI delivery platform helping global organizations like Cushman and Wakefield, Nomura, and Climb Distribution accelerate AI adoption and decision-making at scale. So with that, Larissa, welcome to the event.
Larissa Schneider (00:32.268)
Thank you so much for having me today. My name's Larissa, I'm the co-founder and CEO of Unframe. Really excited to talk to you about the trends that we are seeing in enterprise AI right now. The reason why we ultimately started the platform and the company Unframe was actually we felt since that chat GPT moment, everyone started rushing at the enterprise and taking value from them.
You saw every single SaaS vendor that added AI features trying to increase the annual ticket size of their products. A lot of point solutions started popping up and targeting very narrow use cases at an extremely high price tag. And then there were all of the consultancies offering pure AI strategy sessions with very low impact and high cost. So we were wondering who is actually there to provide value to all the enterprises in the AI world rather than taking it.
Larissa Schneider (01:27.936)
Because what we were seeing was businesses want to adopt AI for all kinds of use cases because of the potential that it has, but they're struggling to really get the business value out of it. Currently, the way that AI works, it requires a lot of tailoring to your specific use case and your enterprise to produce relevant and accurate outcomes. A lot of the point solutions or generic solutions in the market, they don't really work for that.
And something that has 60 % accuracy in AI usually gets 0 % adoption. There are also many enterprise AI use cases that we see that do not have an off-the-shelf solution that's available. And even if there were, think no one at an enterprise level would want to govern and maintain a basket of hundreds of point solutions. And then there were also all of the DIY platforms that popped up recently that
encouraged a lot of teams in-house building their own solutions, but it takes a lot of people and many months to do. So not the ideal way. We also conducted quite a bit of research that I wanted to share here. And I think these points confirm very well what the current market and the landscape in enterprise AI looks like, because it really has arrived in the enterprise. We find that 42 % of large companies have live deployments going on right now. And then there are another 40 or so percent
that are running pilots. But only 1 % of them feel like they have truly matured their AI practice, even though it is the top priority for many of them. The most successful industries that are using AI at the moment are the ones with a very, very strong digital foundation. So you think about your professional services, life science, high tech, telco, and so on. There are
are others like the travel industry, retail manufacturing, as well as the public sector that are still figuring out where AI could best fit their strategy. Other industries like financial services or healthcare, they're investing really heavily. We see a lot of interest in AI from those verticals, but they obviously have to act much slower than some of their peers because of regulation and a lot of legacy systems that they're facing. So in summary,
Larissa Schneider (03:47.348)
a lot of activity, but the scalable impact is still early for most of the enterprise level customers.
So now that we've looked at the industries that are leaning in, let's look a little bit what they're actually trying to achieve and get out of their Gen.AI investments according to a study that was recently conducted by Gartner. In most cases across industries, efficiencies is leading the board. So that is the primary motivation for enterprises to use AI tools right now. Increasing operational efficiency, not really to boost revenue or to cut risks just as much.
Life sciences, banking and investment services are the top examples there. But also here, tailoring really matters because there is very clearly on this chart, no one size fits all kind of use case and every industry is pulling different types of value from AI in their own ways, which means that depending on the solution, you really need to figure out what is the best business value and the...
the domain context to make it work best for these customers. I'll tell you a little bit about a different approach, and then we'll move into some more forward-looking ideas here. Because there isn't just the point solution. There is a different way that we're starting to see in the market that is a lot more customer-centric and helps a lot. And so what you're starting to see pop up, and this is how our managed AI delivery platform also fits into this.
space is a complete AI native way of working nowadays. And the way that we mostly work with our users is actually pretty simple three step process. We usually we discuss and land on the right use case for our customers and use the use case can come from all kinds of different areas can be about observability and reporting or extracting and abstracting off unstructured data to full AI automation and agents.
Larissa Schneider (05:47.725)
We then meet again about five days after this initial conversation with the customer, and we will already give them a complete production ready solution running on our platform that's ready for them to try. And it's not a mock-up, it's not a prototype, it's a real working solution. And in the last step, only when they're happy, we'll move to licensing. So absolutely no upfront cost or commitment until the customer feels like they are getting that business value that they were looking for.
clicked a bit too fast, one more thing here. So how are we actually able to do this? This is actually from the learnings that we had at unframed very early on with our enterprise context that we started working on very quickly. And we learned that even though AI use cases can come from all kinds of departments and industries, the technical components that are required are actually very similar. So we created and collected all of those into a platform of
deep technical building blocks. Metaphor we like to use here is Lego bricks. All of these individual building blocks, they can be tailored or molded to fit the exact use case of the customer. And a way that is being done is with something that we call a blueprint. Blueprint you can think about it as the Lego instructions manual that you see in every single box that helps assemble and configure those into a turnkey solution that's ready to use. And that is extremely high accurate because of the tailored approach that it has.
Once you make it a bit more digestible now and like turn to a specific example with a customer of ours, who's actually interestingly approached us with a really, really great use case, to be honest. And I think one that we can see across many different areas in the world and in different verticals, but it is one of the largest and most influential quality newspaper in the world. And they had an editorial bottleneck. We were able
with our AI solution to cut their proofreading time by almost 70 % and turned what is currently a three-year onboarding process of new employees and team members on the editorial team into immediate efficiencies with our AI solutions. So really help with the efficiencies there and we're able to scale their team and the output of their team without actually scaling the headcount by assisting with AI where it makes most sense.
Larissa Schneider (08:13.896)
And this specific success made us the chosen partner for many more AI transformation projects across the company. But this story is not just singled out. This type of story, it repeats itself constantly in our customer conversations. And we're seeing that the same needs and also the same urgencies impact almost all enterprises. So I want to finish off the session.
with a bit of a look into the future and the predictions that we have for the enterprise AI market. I believe that in 2026, we will start seeing full stack AI companies that pop up. Those are the companies that don't just sell tools, they actually build an entire business powered by AI and then compete directly with incumbents. So think about legacy businesses. They will obviously need to speed up the AI adoption
or they risk being overtaken by these AI native challengers that can move significantly faster and operate a lot leaner than a very legacy business can nowadays.
We also believe that the five-year AI strategy is over. Enterprises that spent years building internal AI platforms with significant resources and investments, they will realize that those efforts didn't actually produce meaningful outcomes. What we will expect to see is them turning to external partners that can deliver business outcomes faster and at a much bigger scale. Because we always say,
If you are an internal developer in a large enterprise company, you will want to work on that company's IP and the core business rather than internal efficiency use cases. If I work for a manufacturer that works on autonomous driving cars, that's what I would want to do as a developer, not so much work on an HR ticketing system or our IT operations processes internally. So assuming that's something that will start popping up.
Larissa Schneider (10:26.006)
We also believe that the quick DIY wave will start to fade out. We've seen a lot of fast, low-code, no-code app generators that entered the market, looked very attractive, and generated a ton of buzz here over the last couple of years. But most of the outputs are really not production-ready, let alone enterprise-grade with all of the tech stack requirements, security protocols that come with an enterprise deployment.
Many of the internal teams tried it though, which is great. A lot of learnings from it. But they quickly reached their limits and started moving on to something else. You also actually start to see that a lot of those companies that had these approaches to no code app generators, they're starting to hire forward applied engineers, the FDE model. Because the pure DIY motion for many of them did not work as a business model either.
I assume that the buzz there is going to die down a little bit.
We also think that the application layer will become more and more important. What we have seen in recent months is that the LLM improvements are really starting to slow down. The real advantage in AI right now is shifting to how companies apply their AI solutions, not really which models they're going to use. And the most valuable apps we see that enterprises use, they really start with a customer back approach rather than a
model LLM forward, and they really solve the existing real business problems in the enterprise, not just try to push in a tool and hope that it will find some application somewhere in the business.
Larissa Schneider (12:15.372)
And then the outcome-based pricing model. I think this will really become a priority for many because enterprises are tired of all of their AI projects failing. I'm pretty sure many of you have seen the MIT report that stated 95 % of AI projects fail right now. Companies and the AI buyers, they will start to push for pricing that is tied to real business outcomes instead of pure consumptions or seed-based models.
Most vendors that exist in the market, they can't deliver on this very well just yet, but the pressure will keep growing. I'm sure about it because the buyers see the outcome-based pricing model as the closest thing that they currently have on the market to a guaranteed value. And honestly, this is exactly why we build Unframe the way that we did because enterprises need their solutions and their AI strategy as well as implementations today.
and not just hope for some empty promises that they will have in some stage in the future. And I'll leave you with that. Thank you so much for your time today, and I hope you enjoy the rest of the event.
Michael Bernzweig (13:24.525)
Fantastic, and that was a wonderful presentation. We're really looking forward to the Q &A session, which is coming up right after this. So for anyone in the audience that has a few questions or details that you're wondering about, we'll be back in just a few minutes. Type your questions into the...
Q &A on your screen. And one question from me, Larissa, a lot of organizations in the mid-market enterprise space were notably late to the game in terms of AI and a lot of smaller organizations jumped right in.
Do you see a foundational reason for that or is it really just that much more complex for an enterprise to roll out AI?
Larissa Schneider (14:20.588)
think we haven't even seen everything that is about to come, right? Like I think initially people are like, okay, let's see how the market develops, how the offering develops. And they were starting to wait a little bit before actually acting and putting their eggs into one basket. But especially this year, you really see this pressure to act both bottoms up as well as top down, you know, like every single boardroom around the world.
has asked their C-level leaders, like, what are you doing around AI right now? So they don't really, they can't afford to wait much longer anymore. But you also have like the bottoms up approach where everyone in our personal life, we've all used chat GPT and use it probably on a daily, if not hourly basis. But then we walk back into the office and all of a sudden everything is clunky old legacy software that no one really wants to work with anymore, right?
So you get the heat from both sides and that has really accelerated the need to act now. And all they need to do is just find the right partners that help them at that scale.
Michael Bernzweig (15:28.352)
I agree, and I think that's why so many founders and executives are here today taking a day out to figure out and answer that question, what are we doing about AI? So thank you so much. We're looking forward to your time during the Q &A session, and we'll be back in just a moment.
Larissa Schneider (15:39.714)
Thanks.
Larissa Schneider (15:47.053)
Thank you so much.