We've been heads down building AI features for community financial institutions for the past year. I wanted to share where our clients stand today. If you're an FI leader trying to benchmark your own progress, I hope this gives you a helpful reference point for comparing with your own institution.
(Part 2 next week: six months of conversations with ~50 FI CEOs and leadership teams — what they're excited about, what's keeping them up at night.)
Every serious AI conversation we have starts in the same place: governance. Our clients want to know exactly how user data is analyzed, where it goes, and what protections are in place before anything else moves forward. That's the right instinct.
64% of Narmi clients have signed our AI Terms and Conditions, unlocking access to Narmi AI features:
• 91% of digital banking + account opening (“Narmi One”) customers have signed
• 51% of account opening only customers have signed
The most common questions we’re hearing: what PII will AI access, and who owns liability? Questions like these used to kill momentum. Now, they are signs of a healthy, engaged leadership team wanting to move quickly, but also abide by regulations.
• 33% are using at least one AI feature in production
• 23% are using two
Our bar for "using" is real. These are institutions running AI features in production with their businesses and consumers on the other end.
1. Narmi MCP Integration: connecting accounts directly to LLMs like Claude or ChatGPT
2. Narmi AI Messages: drafting secure messages for support ticket responses
3. Narmi AI Marketing Promotions & Audiences: building custom audiences via natural language and automating campaigns
MCP being the #1 feature still surprises people when I mention it. Community financial institutions allowing their businesses (and consumers) to connect their account data to Claude or ChatGPT isn't where most people expected adoption to start, but it reflects something real: business owners are using these tools personally, and want their institutions to keep up.
More to come next week on key observations. I hope this benchmarking exercise is helpful as the industry moves quickly on AI adoption.
