Financial AI Frontier: Cognaize's $18M Funding

0:00The wait is over. Dive into Audible's most anticipated collection, The Best of 2025. featuring top audiobooks, podcasts, and originals across all genres. Our editors have carefully curated this year's must-listens from brilliant hidden gems to the buzziest new releases. Every title in this collection has earned its spot. This is your go-to for the absolute best in 2025 audio entertainment. Whether you love thrillers, romance, or nonfiction, your next favorite listen awaits. Discover why there's more to imagine when you listen at audible.com slash best of the year. What can 160 years of experience teach you about the future?

0:44When it comes to protecting what matters, Pacific Life provides life insurance, retirement income, and employee benefits for people and businesses building a more confident tomorrow. Strategies rooted in strength and backed by experience. Ask a financial professional how Pacific Life can help you today. Pacific Life Insurance Company, Omaha, Nebraska, and in New York. Pacific Life and Annuity, Phoenix, Arizona. Cognize has just announced they've raised$18 million to help train a better LLM for the finance sector. And their idea here is one that they say is actually going to be keeping people in the loop more.

1:23So today on the podcast, we're going to be the money on and the implications we see for this in the future of AI. So to kick this off, I want to say that I believe the fintech industry has seen some really massive evolution recently. I think this is definitely due to, in large part, artificial intelligence and a lot of other emerging tech that has come out. I think you can see things like, you know, neobanks crafting individualized customer experiences to analysts building predictive investment models. And I think the AI has really found some solid footing in the AI sector. So, you know, what's the catch?

2:02I think most of the AI driven operations still largely hinge on structured data, which is leaving an unexplored goldmine of unstructured data that could further streamline a lot of these processes. And so that is really where Cognize comes in. So they are a startup from New York, and they are looking to, you know, ride this wave by pioneering a hybrid approach. So their platform targets essentially the processing of unstructured data for financial AI applications and integrates a human touch to essentially refine the process. You know, of course, they just announced their$18 million in funding.

2:42They're using that to scale their business. They're already serving a lot of major clients. They have some top tier credit rating agencies. They have some insurance companies and they also have several financial service businesses. So the funding round, it was actually led by Aragonetic Ventures. They had Metaplanet and some other undisclosed investors that joined in. And it is essentially set to be used. This new, you know, tranche of funding is going to be used to help increase hiring efforts and research. They're going to be working on product development and some business expansion. So the firm essentially has offices in Germany and Armenia, and it isn't revealing its valuation, but it's, you know, it has assured everyone that it's in the hundreds of millions.

3:28I think that's kind of funny. You know, like they announced how much they're raising, not announcing the valuation. It's curious. Anyways, leading the charge on this is Vahy Edonias. That's the startup's founder, who he is the CTO and the CPO. and he's no stranger to the fintech realm so he actually previously launched a firm that offered a credit investment analytics and also risk management ultimately he that was acquired by moody's which is you know a very big credit analysis company so you know big kudos to him for creating a company that was acquired by someone so big his philosophy that ai can accomplish things humans can't but can never replace human intuition is pretty much ingrained in his company operations.

4:16So I think despite the data that is now available in the finance industry, there's a ton of it. I think it is often only structured data that's actually used. So the goal with this new, you know, with this startup here with Cognize is really addressing this imbalance by developing a deep learning platform trained on financial models and an assortment of 1.3 million documents, which is, you know, they're using SEC filings, ESG documents to, you know, loan applications and a lot of others. And essentially, they're feeding that into it. So the platform essentially assists financial analysts who can correct readings and draw conclusions based on outcomes.

4:57So the CEO is Al Essien. And so he, this is a quote from him, he says, For a bank, there are three objectives. You can try and fail to build AI in-house, or you can use a general AI model like ChatGPT with an army of consultants. The third choice, us. We empower and educate you. So he is an entrepreneur with a history of a bunch of successful exits. He exited from VMware, and then he joined Cognizize shortly after it was started. Once his aerial imagery analytics firm, Intelinair, found its successor. so while tech giants like open ai google and anthropic are making some i would say fairly significant strides in general ai um smaller and more specialized players like this i think are going to be able to find a real niche in the market that is going to appeal to people um you know cognize for banking and investor type people but i think this is a trend we'll definitely see in the future where we have these general models that can do a lot that we're going to see a lot of very specific models that are going to be really successful um i think there'll be thousands of them i think it'll be a very fragmented field um and that's really the way that you get the best results and so specific niches will know specific models and use them um religiously because they just do so much better than a general model at what they're trying to do so they actually said um or one of their investors who is viken doozjan from ergonotic ventures he said we're excited to join forces with cognize as they harness ai and large language models to revolutionize finance he also praised their ability to derive insights from large amounts of unstructured financial data he said remarkable precision and speed and really he thinks that this is going to lead to better decision making risk assessment and the discovery of patterns which previously were hidden by complexity and also just human error so metaplanet's managing partner which is roux mijad he also has enthusiasm.

7:00He said that their inevitable growth and its ability to deliver repeatable, measurable value through AI and finance was really exciting. And so as much as AI has really disrupted, I think, the financial landscape, I think that this company really believes that there's going to be room for a really kind of nimble, focused enterprise player like Cognize. And they also do acknowledge the importance of really leveraging broad spectrum solutions like chat chippie tea where it aligns with the business's objectives right not every business is going to need a highly special a specialized tool a lot of them will be able to use kind of these broader tools and maybe integrating both actually makes sense in a lot of cases so um they do say you know we're more agile which gives us an edge however it's only the paranoid who survive so we're also leveraging tools like chat chippie tea where it makes sense it looks like they're like Like, you know, you hate to say hedging your bets when an AI company or when any company is not just exclusively using their own tools because, you know, it's like, is it really hedging their bets?

8:04They're definitely making a great piece of a great LLM in the finance sector. But at the end of the day, ChatGPT does a lot of really general things great. And so I think there's no harm in integrating that on top of their own thing. Now, of course, if they were exclusively just a rapper on top of ChatGPT, I don't think they could have raised that much money. So they really do need to put a big emphasis on the fact that their own LLM is going to be really powerful in accomplishing the task that it's set to do. But for a lot of the general stuff, ChatGPT makes sense. And to be honest, this actually makes a lot of sense.

8:33That kind of strategy or that kind of concept, I think it's important. We're going to see it. It's going to be a big trend in AI. And I think it actually gets down to OpenAI and how they actually have ChatGPT doing tasks. So recently, the model weights and a bunch of other things from GPT-4 were leaked. And it was found that they use what are called, I think, experts. And so they have 16 experts within GPT-4 that essentially are models or, you know, software trained to do specific things very well. And so when GPT-4 gets a question, it shoots it over, has an AI that decides which of these 16 experts can respond to it best.

9:10It shoots it over there. It's got like multiple layers that kind of hit the text and edit it. And that's how the response has got so much better. and so following that same kind of concept with this I think using creating any sort of tool so let's say this is finance tool someone asks them a financial question it can be fed into chat should be t to determine what kind of question it is what is going to answer its question best and then chat should be t could spit it over to their own model that's going to really do some of the heavy data crunching and you know algorithmic detection or whatever their their financial tool does.

9:43But I think leveraging ChatGPT and integrating the two together honestly is still a good play because otherwise they would literally have to try to recreate ChatGPT in a sense and also, you know, build their own. So I think a double integration like this makes sense. It's the same pattern ChatGPT and OpenAI are using. It's just they're able to really refine it. And it's, you know, it's going to be the equivalent of instead of having 16 experts in GPT-4, there's going to be 10 000 experts and all of these people might layer themselves like with openai and chat gpt to get like the first layer send it over to them and they're gonna uh they'll get more specific answers so this is a really interesting area we'll follow in the future um an interesting company it's left to be said you know how they do they raised 18 million dollars really impressive so i'll be excited to follow this in the future.