The First AI Score: Where Credit Unions Stand Now

This is the first pass of the AI Score, and it sets the baseline everything that follows will be measured against. As of September 23, 2026, 4,299 credit unions have been scored. Public evidence of AI work was found for 272 of them: 3 at five stars, 97 at four, 169 at three and 3 at two. The remaining 4,027 sit at one star, which means no public evidence was found in this pass. Because this is the first scoring date, no credit union moved up this time; there was nothing yet to move up from.

Three credit unions reached five stars, the level for agents that act with human approval. Baxter CU in Vernon Hills, Illinois, with $6.4 billion in assets and 372,400 members, is described by salesforce.com: "they are extending Agentforce into the product sales lifecycle with lead nurturing agents, with plans to launch multiple new agents each quarter." Del-One FCU in Dover, Delaware, with $983 million in assets and 106,500 members, draws this from delawarelive.com: "The credit union has deployed an autonomous digital assistant named Valeria to review documents submitted by automobile dealerships through Del-One's indirect lending program." And SkyOne Federal Credit Union in Hawthorne, California, with $863 million in assets and 58,940 members, is quoted in prnewswire.com: "By integrating AI-driven, role-based digital agents, we are not only improving operational efficiency within our Credit Union." Two of the three sit in the $250 million to $1 billion band, a reminder that this level of work is not reserved for the largest balance sheets.

The four-star group, where AI is embedded in workflows, numbers 97. Its largest members include Suncoast CU in Tampa, Florida, at $20.8 billion and 1,408,000 members; Alliant CU in Chicago at $20.0 billion, where thefinancialbrand.com reports that "Part of that strategy is unfolding through a partnership with Upstart, a digital lending matchmaker driven by artificial intelligence decisionmaking"; SCCU in Melbourne, Florida, at $9.2 billion; Redstone FCU in Huntsville, Alabama, at $8.6 billion; HVCU in Poughkeepsie, New York, at $8.1 billion, where posh.ai notes that "With Knowledge Assistant deployed to help agents with customer calls in real time, the credit union achieved a first-call resolution (FCR) rate of 89%"; and Gesa to Go in Richland, Washington, at $6.8 billion.

Evidence is denser at the larger end. Among the 92 credit unions above $5 billion, 48 have public evidence, a 52% share. In the $1 billion to $5 billion band, 111 of 387 do, or 29%. The $250 million to $1 billion band accounts for 91 of 671, at 14%, and it is also home to two of the three five-star names. Smaller bands show less so far: 18 of 1,243 between $50 million and $250 million, and 4 of 1,906 below $50 million.

The score runs one to five stars, with one star meaning no public evidence was found, two employee copilots, three AI connected to a credit union's own data, four AI embedded in workflows and five agents that act with human approval; every star above the first rests on a linked public source with a verbatim quote, scores never fall, and the next check comes in two weeks.