The news: BNY will integrate Google Gemini Enterprise into Eliza, its in-house AI platform. Gemini will give Eliza deep research tools and let BNY’s employees build AI agents that draw from and act on the bank’s vast libraries of financial data.

Zoom out: 70% of bankers in an American Banker survey recently said that agentic AI will “have a significant or game-changing impact on the banking industry,” but only 18% said their bank was “actively deploying” agentic AI, and 33% were “testing through pilot(s).” Being in production with leading-edge AI tools creates a huge competitive lead: Advances in AI have exponential benefits based on how quickly the technology evolves.

But contracting with vendors for genAI tools like Gemini is only a half measure. JPMorgan Chase, which one benchmark says is leading the AI race, spends $2 billion per year on AI alone among $18 billion spent annually on technology, which includes substantial investments in talent and infrastructure. Royal Bank of Canada, which has been ranked highly on AI maturity, has an internally developed AI research group. Bank of America is pouring money into enterprise-scale AI.

What’s next: The haves and have nots of banking’s AI era are coming into relief.

The top strata are institutions that have invested heavily in data infrastructure that supports AI’s needs and in R&D that competitors can’t buy. The next strata are institutions that are large and sophisticated enough to develop a dedicated AI strategy and execute on it but aren’t building as much in-house. Banks in the final strata haven’t found ways to use AI at scale but need to now.