UBS, Switzerland’s largest bank, which took over Credit Suisse in 2023, has 608 artificial intelligence (AI) use cases in production, 81% more than a year earlier. More than 1,090 other applications are in development, and its employees run more than 12,700 personal AI agents. These figures accompany the results for the third quarter of 2026, published on 28 October 2026, according to Fintechnews Switzerland on 2 November 2026. They offer a rare measure of AI in banking at the scale of a large Swiss institution.
What do UBS’s AI figures cover?
The figures describe the state of deployment in the third quarter of 2026, from July to September. Fintechnews Switzerland reports the following figures:
- 608 AI use cases in production, up 81% year on year;
- more than 1,090 applications under development;
- nine major so-called “end-to-end” AI initiatives;
- more than 12,700 personal AI agents used by employees;
- more than 18,000 developers using AI tools to write software.
One last figure concerns a specific tool. Since the start of 2026, the internal tool STAAT has produced more than 30 million analyses, passed on to teams of financial advisors. This is a volume of generated recommendations, not a measure of their use by advisors or of their effect for clients.
These figures are published by the bank itself. Those reported by Fintechnews Switzerland give neither the definition of a “use case” or a “personal agent”, nor the share of tools aimed at clients rather than at internal work.
How has AI use evolved at UBS?
The growth over one year is strong. According to Evident Insights, the platform that publishes the Evident AI Index, UBS had launched more than 300 AI use cases in 2025. This figure was reported by 2clics on 21 October 2026, at the time of the Evident AI maturity ranking, in which UBS ranks sixth worldwide and remains the only European bank in the top 10.
The deployment also relies on training. According to the same study, more than 49,000 employees have enrolled in online AI courses since 2024. In July 2026, more than 21,000 people, or 28% of the workforce, had completed the internal “UBS AI Citizen” programme.
What does UBS expect from artificial intelligence?
The bank puts forward three objectives. According to Fintechnews Switzerland, UBS says it is investing in technology, including AI, to improve the client experience, strengthen its infrastructure and become more efficient.
The announced gains remain isolated and self-reported. For its legal department, UBS uses Harvey, an AI platform that drafts, summarises and reviews legal documents. According to the bank, quoted in the Evident study, it saves at least 1 million dollars, and 1 million more in 2026 by bringing back in-house work that had been outsourced. No overall figure for AI-related savings accompanies the 608 use cases.
The case of the 18,000 developers calls for a caveat. A Harvard-affiliated study, presented by 2clics on 18 October 2026, measures 20 to 30% more code with AI, without a significant increase in finished software. Human code review becomes the bottleneck: the review time rises from 7.03 to 10.48 days on average. There is no indication of whether UBS observes the same effect.
UBS compared with other Swiss institutions: how wide is the gap?
UBS stands well above the sector average. The Swiss Financial Market Supervisory Authority (FINMA) surveyed around 400 authorised institutions between the end of November 2024 and mid-January 2025. According to its press release of 24 April 2025, around half were using AI or developing their first applications.
| Indicator | UBS (third quarter of 2026) | Institutions surveyed by FINMA (end of 2024 to early 2025) |
|---|---|---|
| AI applications in use | 608 in production | around 5 on average |
| Applications in development | more than 1,090 | around 9 on average |
| Source | UBS, via Fintechnews Switzerland | FINMA survey, press release of 24 April 2025 |
The comparison is indicative. The periods differ, and the FINMA average covers only respondents that use AI, banks, insurers and asset managers combined. According to FINMA, 91% of these users rely on generative AI, and the smallest institutions often rely on applications developed externally. For an SME, the challenge resembles the one described in an OpenAI report on AI agents in SMEs in Europe: cost, suitable tools and skills.
What framework does FINMA set for AI in banking?
There is no AI-specific rule for Swiss banks. On 18 December 2024, FINMA published a guidance on governance and risk management relating to AI. In it, FINMA lists the risks: model risks (lack of robustness, accuracy, explainability, bias), data-related risks, IT and cyber risks, dependence on third parties, and legal and reputational risks.
This text describes expectations, not a new obligation. FINMA noted in it that most institutions were still at an early stage, with governance still being built. Its principle, restated in April 2025, is “same business, same risks, same rules”: the same activity calls for the same rules, whatever the technology. FINMA also recommends that institutions contact it in good time if they plan to use AI in critical processes. The page FINMA and artificial intelligence details the scope of this guidance.
Our analysis: for a client in Switzerland, the bank remains their point of contact, whatever AI tool is used. With 12,700 personal agents, the question of controlling each of them becomes concrete for UBS as well as for its supervisory authority.
Is AI changing employment at UBS?
The results documents do not link AI to employment. The report for the third quarter of 2026, filed on 28 October 2026 with the Securities and Exchange Commission (SEC), the US securities regulator, counts 97,593 internal full-time equivalents as at 30 September 2026, compared with 104,427 as at 30 September 2025.
This decline is taking place during the integration of Credit Suisse, which the report describes in detail. As at 30 September 2026, around 98% of the former Credit Suisse IT applications to be decommissioned were no longer in use. According to a previous 2clics article on UBS’s third-quarter results, the bank has cut around 20,000 jobs worldwide since 2023 and has 34,000 in Switzerland. The report does not attribute any of these cuts to AI. On AI employment in Switzerland beyond UBS, 2clics has published a separate analysis.
Key figures
- 608 AI use cases in production at UBS in the third quarter of 2026, +81% year on year;
- more than 1,090 AI applications in development;
- more than 12,700 personal AI agents used by employees;
- more than 18,000 developers equipped with AI tools;
- more than 30 million analyses produced by the STAAT tool since the start of 2026;
- around 400 institutions surveyed by FINMA at the end of 2024 and early 2025, of which around half were using AI or preparing to do so.
What we don’t know yet
- The share of the 608 use cases that directly affects clients, and the share that serves internal work.
- The total savings UBS attributes to AI: the bank only cites isolated gains, such as those from Harvey.
- How the 12,700 personal agents are overseen and controlled.
Sources
- Fintechnews Switzerland: “UBS Grows Live AI Use Cases to 608 as Staff Run 12,700 Agents” (2 November 2026)
- UBS Group AG: report for the third quarter of 2026, Form 6-K filed with the SEC (28 October 2026)
- FINMA: press release “FINMA guidance on governance and risk management when using artificial intelligence” (18 December 2024)
- FINMA: press release “FINMA survey: artificial intelligence gaining traction at Swiss financial institutions” (24 April 2025)
- 2clics.com articles cited in the text (Evident ranking, FINMA framework, UBS’s third-quarter 2026 results, code review, AI agents in Europe)
