Evident AI: Tracking How Banks Put AI to Work

Banks announced more AI use cases in the second quarter of 2026 than in any quarter on record, and the leaders have moved past asking whether AI works. Now they’re asking how to scale it. Few organizations track that shift as closely as Evident AI. This fall, interop.io was a participating partner at Evident’s AI Symposium in New York, and this week we’re featured in the Use Case Corner of its newsletter, The Brief. As the interoperability layer behind some of the most advanced desktops in financial services, we see that shift up close, including at many of the banks ranked on the Evident AI Index.

What is Evident AI?

Evident is an intelligence platform that measures and tracks how financial services firms are adopting AI. Its flagship Evident AI Index benchmarks major banks and insurers across more than 70 metrics, giving leaders a clear view of how their AI capabilities, use cases and returns compare with their peers.

Eight interop.io clients are ranked in the Index, including the #1.

Evident also runs the AI Use Case Tracker, a library of more than 500 AI use cases announced by the world’s largest banks and insurers, categorized by business line, function and end user. Its latest quarterly analysis found that wealth management and commercial banking use cases tripled as AI moved beyond retail banking.

Founded by Alexandra Mousavizadeh and Annabel Ayles, the company set out to bring transparency to a market full of AI hype. Beyond the Index and the Tracker, Evident publishes research on AI ROI and talent across the sector, runs a members-only community for AI leaders.

Why It Matters to Us

Evident’s research shows banks moving from AI pilots to AI in production. In our work with clients, the hard part of that shift is rarely the model. It’s getting AI capabilities into the user’s actual workflow, alongside the CRM, portfolio, research and trading tools they already rely on, so that people use them every day.

That’s the problem io.Connect solves. As the interoperability layer on the desktop, it lets AI assistants and agents share context with the applications around them, so a recommendation shows up next to the right client, the right position or the right workflow, instead of in yet another window.

Use Case Corner in The Brief

This week’s issue of The Brief includes interop.io in its Use Case Corner, which looks at what it takes to bring dozens of AI capabilities into a single advisor desktop, and what happens to adoption when you do. What do nearly 4 million advisor questions answered, more than 15,000 users and active engagement above 80% look like for a large investment bank? If you’re building AI into your own platforms, it’s worth a read. The AI-Centric Desktop for Wealth Advisors.

Evident AI Symposium

On October 22, Evident held its fifth AI Symposium in New York. The day brought together technology and AI leaders from across banking and insurance, including Goldman Sachs, Bank of America, Capital One alongside headline sponsors Anthropic, EY, Synechron and Akka.

Evident’s co-founders opened with a clear message: after four years of investment, leading banks are past experimentation and are now scaling what works. A few themes came up again and again:

• Controls make scale possible. Speaker after speaker argued that AI agents need named owners, their own identities, audit trails and a way to shut them down, and that these controls should be built once and reused across every deployment.
• Cost discipline is the new priority. The era of broad access and maximum token use is giving way to matching models to tasks, using smaller fine-tuned models where they fit, and tying AI spend to measurable business value.
• Evaluations are becoming infrastructure. Banks are using evals to set baselines, prove productivity gains and shorten model risk reviews, and some panelists described strong evals as a lasting competitive advantage.
• Watch what agents do, not just what they say. Several firms described monitoring AI agents in live workflows, with people and systems ready to step in, as essential to scaling with confidence.

Those themes line up closely with what we hear from clients. The model is only one part of the system. What decides whether AI gets adopted is the shared infrastructure around it: governance, visibility, and getting AI into the workflows people already use.

Evident has posted summaries and full recordings of every session. Explore the Symposium session recaps. We hope to see you at the next Evident AI Symposium!