Relationship banking is usually based on people talking to people — but a newly founded fintech company wants to introduce agentic AI tools to help smaller firms scale their relationship banking services.
David Sosna, founder and former CEO of
Sosna's emphasis is on serving regional banks with agentic AI technology, particularly in their relationship development with small and medium sized businesses. Regional banks, according to Sosna, could have a competitive advantage against big banks by investing in their relationship banking strategy.
"In the next few years, many more banks will understand that they have to play on their strengths if they want to keep their identity," Sosna said. "There's a lot of banks in the United States. They're fighting for every customer, and they have to define themselves in a way that will deliver a unique set of capabilities for their customers. I do think that relationship banking, if you can scale it correctly, will be a significant competitive edge for these banks."
The challenge for smaller banks, according to Sosna, is scaling relationship management across a wider customer base due to limited resources.
"The problem that they have is that they don't have enough high-quality bankers to cover all customers," he said. "They have a scale issue, and a lot of these bankers are young. If you think about small businesses and other customer segments, there's hundreds of them assigned to the same banker. It's not really something that they can capitalize on."
Sympera AI's agentic tools are designed to be used by a financial institution's relationship managers. David Hall, a consultant to big banks who is working with Sympera AI as an independent advisor, emphasized that Sympera's AI tools will not replace a relationship manager's role but instead help them do their jobs better.
"Across the board, what the AI does is it really just helps humans be better at the work that humans do," Hall said. "Whether it's, 'Help me do this at a greater volume. Give me guidance on what might be the next thing I should talk to the customer about, because AI can analyze large volumes of data much faster than I can. Then let me actually have it do some of the initial work.' The beauty is that Sympera can cross all those functions, but within any bank."
Although Sosna is focusing on regional and community banks as he launches his new venture, Hall sees potential for agentic AI to enhance relationship banking across the financial services sector regardless of the size of the bank.
"Banks of all sizes want to be able to support small businesses in a way that it's profitable for the bank," Hall said. "That's always been a challenge, because small businesses are generally not higher revenue generators at an individual client level. Being able to scale the services and the support and knowing, 'I'm going to reach out to these customers about these exact things because I can see, with the help of AI, that they have or are going to have a need there.' That helps bankers hone in, be more effective and make those interactions more profitable."
The technology behind agentic AI is still in development, however. The potential for problems such as AI agents performing errors at scale, and thus liability concerns if those errors cause money losses for banks or customers, has caused some banks to
Sosna believes that focusing on the agentic side of AI could reduce the risk of errors that generative AI and LLMs are currently known for.
"If you go only after an LLM as an LLM, the chances of errors and hallucinations is very high because you're asking ChatGPT to do it all," Sosna said. "When you split the problem into designated tools you can control them, you can verify them, and so then the error rate goes down dramatically. That's why we are so excited about agentic flow, which means small agents doing big things together and reducing the number of errors."
