Banks raced on AI in 2026. Only 1% of their use cases show a dollar return.
The Evident AI Index, the most-cited outside ranking of how banks use AI, published its 2026 edition on October 6. The leaderboard barely moved, the pace of adoption jumped, and the evidence that AI is paying off remains thin outside a handful of banks.
Banks are among the biggest corporate buyers of AI, and AI in trading desks, research and risk is the part of the technology closest to markets. Evident, a research firm, scores 50 of the largest banks in North America, Europe and Asia-Pacific on 75 indicators across four pillars: talent, innovation, leadership and transparency. The index is built from public data, which makes it an outside-in view rather than an audit.
The ranking
JPMorganChase ranks first for the fifth time, placing first or second in all four pillars. Capital One is second and RBC third, the top non-US bank. Commonwealth Bank of Australia (CommBank) is fourth, and Wells Fargo moved into the top five. Every bank in this year's top ten has placed there before.
| Rank | Bank | Note |
|---|---|---|
| 1 | JPMorganChase | Fifth year at the top |
| 2 | Capital One | First in talent and innovation |
| 3 | RBC | Top non-US bank |
| 4 | CommBank | Unchanged |
| 8 | Citi | Up four places |
| 10 | TD Bank | Up three places |
| 36 | Westpac | Up nine places |
Faster adoption
Evident reports that the average bank's score rose 26%, or 9.6 points, year over year, nearly three times the moving average of 2023 to 2025 Context: n=50 banks, 75 indicators · 2026 index, published Oct 6, 2026 · Evident AI Index · margin: not disclosed. It calls this the fastest acceleration since the index launched. The hiring data points the same way: AI enablement staff, the people who turn models into working tools inside the business, grew 21%; AI product management staff grew 10%; model risk staff grew nearly 10%.
The return gap
The weak spot is proof. Of the AI use cases Evident tracks, only 12% report impact against operational measures, and about 1% disclose concrete financial returns Context: n=more than 1,100 use cases analyzed · 2026 index · Evident AI Index Key Findings · margin: not disclosed. Evident says it analyzed more than 1,100 use cases for this edition. The top ten banks account for 25% of use cases with impact claims.
At the group level, 12 banks now report a realized or projected return on their AI activities, up from 8 a year earlier. CommBank and TD Bank set their first targets this year. RBC is working toward about $700 million of AI value it has promised investors by next year, and TD says it has made $137 million from AI so far this year, according to Evident's review.
The trading desk example
The most market-relevant case in Evident's review is JPMorganChase's SpectrumIQ, a system that combines research, market data and risk information to support buy and sell decisions and then execute trades in the bank's asset management arm. Evident reports that nearly 75% of equity trading and 85% of FX trading in that arm is now automated, and that equity automation has saved clients $4 billion in trading costs since inception Context: n: not disclosed · since inception to 2026 · JPMorganChase disclosures, via Evident · margin: not disclosed. These figures come from the bank's own disclosures, not from independent measurement.
Controls are splitting the field
Evident finds that 80% of leading banks have adopted sophisticated control mechanisms for AI, such as automated guardrails and post-deployment monitoring, against 40% across the rest of the industry Context: n=50 banks, 75 indicators · 2026 index, published Oct 6, 2026 · Evident AI Index · margin: not disclosed. As banks move from chatbots to agents that act on their own, those controls become the bottleneck for scaling.
What it means outside banking
Banks are a test case for the question hanging over all AI spending: does it pay? The 2026 index says adoption is accelerating and a few leaders can show returns, but most of the industry still cannot put a number on it. That gap is the same one investors are weighing when they price the companies selling AI to banks.
What it shows / What it doesn't
It shows
- How 50 large banks rank on public evidence of AI adoption in 2026, and how fast the average score rose.
- How few AI use cases come with disclosed financial returns.
- Examples of AI value claims made by individual banks, as compiled by Evident.
It doesn't show
- Independent verification of banks' AI savings or revenue claims, which come from the banks themselves.
- Internal AI work that banks do not disclose publicly.
- Whether AI spending will raise any bank's profits or share price.
- The full methodology, which sits in Evident's downloadable report.
Why it matters
AI trading tools at banks are the institutional version of what retail traders are trying with AI. The index shows rapid deployment but scarce proof of payoff, a useful reality check on AI return claims anywhere in finance, including in tools marketed to individual traders.
Primary sources
- Verified2026 Evident AI Index | Banks: Key Findings Report (Evident Insights)Score growth, use-case return gap, ROI reporting, controls, published October 6, 2026.
- VerifiedInside the 2026 AI Index (Banking Brief, October 8, 2026) (Evident Insights)Ranks, talent growth, SpectrumIQ, RBC and TD figures.
- LinkedEvident AI Index for Banks, ranking and methodology (Evident Insights)Full ranking page, not re-checked line by line.
In our directories
How we used AI: We read Evident's 2026 key findings page and its October 8 briefing in full and kept only figures Evident states directly. Bank-level claims are attributed to the banks via Evident.
Not financial advice. For information and education only. AI model outputs and research results can be wrong and do not predict future returns. Nothing here is a recommendation to buy or sell any asset. See our disclaimer and methodology.