AI has had a major impact on the financial sector over the years, helping neobanks provide personalized service to customers, lenders evaluate loan applications, digital providers detect fraud and security issues, analysts run investment forecasting models, and more.
However, most of the work being done today is in the domain of structured data.Along the way, there’s a wave of unstructured data waiting to be mined and used, says a New York startup Konieze A mixed approach is being taken. It built a platform for processing unstructured data for financial AI applications and complemented the platform with “humans in the loop” to complete the work.
Today, the company announced $18 million in funding to expand its business and land some big clients for its services. Clients include two of the three major credit rating agencies, major insurance companies and financial services firms.
Argonautic Ventures led the round, with participation from Metaplanet and other unnamed investors. Cognaize did not disclose its valuation, other than to confirm that it is in the hundreds of millions of dollars.
Cognaize also has offices in Germany and Armenia, and the funding will be used for recruitment, research, product development and business development.
Cognaize’s founder, Vahe Andonians, who is also the chief technology officer and chief procurement officer, previously founded another fintech company that provided analytics and risk management around credit investing, which was eventually acquired by Moody’s. His approach there, and Cognaize’s, is based on the idea that AI might be able to do things humans can’t, but it still can’t replace humans.
Cognaize takes the premise that while the financial industry has access to a seemingly limitless amount of data today to better understand its services, market conditions, and customers, it typically only uses a fraction of it. Data – structured part.
The startup has built a platform using deep learning trained specifically on financial models and a wide variety of documents (1.3 million in total), which may contain many different “units” of information that require a more specialized eye to “read” “. (These documents cover loan applications and also include SEC filings, ESG filings, presentations, fiduciary reports, etc.)
In turn, the platform is used by human workers (often financial analysts) to help correct what is being read, and to make conclusions and decisions based not on the results.
“If you’re a bank, you now have three options,” said Cognaize CEO Al Eisaian. “You can try to build AI capabilities in-house, but forget about that. You can use ChatGPT to go through a generic AI model and try to implement it with a team of consultants. Or option three is us. We help and educate you.”
Eisaian, a repeat founder with exits from companies like VMware, was not a founder of Cognaize but joined Cognaize shortly after it started. The reason for the delay was that he needed to find a successor at the last company he founded and led, an aerial image analysis specialist called Intelinair.
The development of startups like Cognaize in AI highlights an important theme in the field: while there may be many companies, such as OpenAI, Google, Anthropic, etc. Building “Large” In the evolution of large language models, there is an equally interesting trend of very strong players focusing on specific domains and use cases. These players may still be building “big” LLMs, but they focus more on depth of scope than breadth.
Yes, the biggest of them will likely try both, but experts may always be able to speak the language of their customers more directly, and that’s what investors are probably betting on too.
“We are excited to partner with Cognaize as they apply the transformative power of artificial intelligence and large language models (LLM) to the financial sector,” said Viken Douzdjian, managing partner at Argonautic Ventures, in a statement. industry, but large volumes of unstructured financial data create a myriad of use cases that require finance-specific generative artificial intelligence. The Cognaize platform can process large volumes of unstructured financial data and extract insights with extreme precision and speed to enhance decision making , risk assessment, and uncover patterns and trends that were previously obscured by complexity and human error. We strongly believe that Al, Vahe, and the Cognaize team can define how the financial industry interacts with AI.”
“Cognaize is a company to watch as they are one of the first in the financial industry to deliver repeatable and measurable value through artificial intelligence. Investing in Al, Vahe and the entire Cognaize team was an easy decision.” Metaplanet Managing Partner Rauno Miljand added. “They have harnessed the power of artificial intelligence, as evidenced by the enviable growth of the Cognaize business, the global financial leader they have become for their clients, and their unparalleled technology roadmap. They are rapidly redefining how the financial industry leverages Modern AI is here to harness the power of its own data to slash costs while creating new competitive advantages.”
Of course, the most compelling argument for a more targeted approach is that they lead to better results and are trained to the specific needs of the firm; moreover, given the smaller parameters of the LLM, less computing power is required , they may be less expensive to run.
“There’s always an opportunity because we’re more agile and focused,” Andonians said. “That gives us an advantage.”
“Having said that, only the paranoid survive, so we’re also leveraging things like ChatGPT where it makes sense,” he added after a pause.
