🔍 Read the full analysis: How UBS’s 12,700 Personal AI Agents Fit Into Bank Operations on ThorstenMeyerAI.com
TL;DR
UBS reported on October 28 that employees were running more than 12,700 AI personal agents, alongside 608 AI applications in production and more than 1,090 in development. The bank has not disclosed what the agents can access or do, or how they are reviewed; its CEO said UBS does not allow outside models to act for clients.
UBS said employees were running more than 12,700 personal AI agents to support their daily work, as the Swiss bank outlined its AI activity in third-quarter results released October 28. The figure sits alongside 608 AI use cases in production and more than 1,090 in development, but UBS did not explain what the personal agents can access, whether they can take actions, or how the bank oversees them.
In the results release, UBS also reported nine large-scale, end-to-end AI initiatives and more than 18,000 software developers using AI tools in development work. The bank said those tools had led to measurable productivity gains, but did not provide a figure for the gains. It also said its STAAT tool had generated more than 30 million insights so far in 2026 for its U.S. financial-adviser teams.
The bank counted 608 live AI use cases, an 81% increase from a year earlier, and more than 1,090 additional applications in development. The figures are UBS’s own disclosures; Reuters and PYMNTS also reported the 12,700-agent figure, citing the results. The release does not break down the agents by business area or say how many employees use them.
UBS had said in its July 29 second-quarter release that all employees could create their own AI agents. The bank described that access as a way to deploy AI safely and at scale while allowing staff to spend more time on judgment, collaboration and client value. That statement establishes availability, not how many staff members created or actively use agents.
Infographic 01 · What UBS reported
12,700 personal AI agents, inside one bank
The AI figures from UBS’s third-quarter results, and how fast they grew in three months.
From permission to 12,700 agents
UBS 2Q26 release (29 Jul 2026) vs. 3Q26 release (28 Oct 2026)
What the agent number is not
- Not 12,700 people. One employee can run several agents; UBS does not say how many people use them.
- Not a definition. UBS does not say what an agent can access or do on its own.
- Not a result. “Measurable productivity gains”, but no hours or costs published.
- Scale, for context: about 13 agents per 100 internal full-time staff (12,700 ÷ 97,593, own calculation).
“Our employees are running over 12,700 AI personal agents to support them in their daily activities.”UBS, third-quarter 2026 results release, 28 October 2026
Also in the release: 9 large-scale AI initiatives and more than 30 million insights from the STAAT tool for US financial advisor teams so far this year. UBS placed 6th of 50 banks in the 2026 Evident AI Index.
Sources: UBS 3Q26 release, 28 Oct 2026 · UBS 2Q26 release, 29 Jul 2026. Reported identically by Reuters, 28 Oct 2026 and PYMNTS (29 Oct 2026).
Powered by Thorsten Meyer AI
Oversight Becomes the Operational Test
The scale UBS disclosed makes governance and access controls central questions for banks considering employee-built AI tools. A system that drafts a summary for a worker is different from one that sends a message, changes a record or approves a transaction. The results do not say which category UBS’s personal agents fall into, so their number alone cannot show the extent of operational or client risk.
The disclosure also offers a concrete example of AI adoption inside a major regulated institution: alongside agents, UBS reports hundreds of live applications and thousands in development. That matters to other companies weighing staff access to agent-building tools. The practical issue is not only whether employees can use them, but how an organisation tracks what has been created, the data each tool can reach and the actions it can perform.
UBS’s reported staffing figures do not establish that AI has replaced employees. The bank’s workforce fell by about 1,500 full-time equivalents during the quarter, to 97,593, and UBS and Reuters attributed the decline to the Credit Suisse integration. UBS did not link that reduction to AI, and the release provides no basis for drawing a causal connection.
Popular right now
De'Longhi 15-Bar Pump Espresso Machine with Milk Frother & Temperature Control
Dyson Ball Animal 3 Upright Vacuum
Samsung 85-Inch QN70H Neo QLED 4K Smart TV (2026)
As an affiliate, we earn on qualifying purchases.
From Employee Access to Bank Controls
The 12,700 figure is part of a short AI update in UBS’s third-quarter results, a release mainly focused on financial results, capital requirements and the Credit Suisse integration. The bank’s July statement that all employees could create agents preceded the October count by one quarter. Over that period, UBS’s reported live use cases rose from 560 to 608, while applications in development increased from more than 920 to more than 1,090.
UBS has described other employee-facing tools, including an in-house assistant called Red, which its technology materials say was rolled out to around 100,000 employees, and an AI-powered HR assistant for tasks such as checking policies and booking leave in Microsoft Teams. These examples provide a sense of workplace AI at the bank, but UBS has not said that either product accounts for the personal-agent total.
A separate boundary concerns AI acting for customers. At an October 27 event at ETH Zurich, CEO Sergio Ermotti said UBS does not allow outside AI models to enter the bank and act on behalf of clients, such as managing accounts or moving deposits, according to Reuters. That statement addresses external models acting for clients; it does not explain the permissions granted to employee-created agents.
„Employees are running over 12,700 AI personal agents to support them in their daily activities.“
— UBS, in its third-quarter 2026 results release
Agent Permissions Remain Undisclosed
UBS has not defined what it counts as a personal AI agent, identified the models or platforms behind them, or described the systems and information they can reach. It is also unclear whether they only prepare drafts and recommendations or can carry out tasks without a person taking a further step.
The bank has not disclosed how agents are tested, monitored or approved, how many employees use them, or whether any of its 608 production use cases face clients. Its results say developers using AI tools recorded measurable productivity gains but give no metric or method for calculating them. The available figures therefore show reported deployment, not the agents‘ effectiveness or risk profile.
Reuters reported Ermotti’s restriction on outside models acting for clients, but that does not settle the rules for internal employee agents. The source material also cites Swiss regulator FINMA’s expectation that banks maintain an AI inventory with risk classifications, while noting that wider access to applications makes inventory management harder. UBS’s specific inventory and control procedures were not detailed in the results.
More Detail Needed on Controls
The next useful information would be a description from UBS of how it defines an agent, what permissions agents receive and what review is required before they are used. Further reporting or disclosures could also clarify whether agents can take actions, how the bank monitors their outputs, and whether any are used in work directly affecting clients.
UBS has not announced a specific follow-up date for those details. The 12,700 figure is a snapshot reported in the October 28 results, while the bank’s wider AI deployment continues to be represented by its counts of live and in-development applications. Until further information is provided, the confirmed scope is employee use at the scale UBS reported, not autonomous operation across client accounts.
Key Questions
What did UBS report about its AI agents?
UBS said in its October 28, 2026, results release that employees were running more than 12,700 personal AI agents to support daily activities.
What can the 12,700 agents do?
UBS has not specified their capabilities. The results do not say what data or systems agents can access, or whether they can take actions independently rather than prepare work for employees.
Are UBS’s AI agents allowed to manage client accounts?
Ermotti said UBS does not allow outside AI models to act for clients, such as managing accounts or moving deposits. That comment does not explain the permissions of employee-created agents.
Did AI cause UBS’s workforce to shrink?
The source material reports that UBS’s workforce fell by about 1,500 full-time equivalents in the quarter, to 97,593, and attributes the reduction to the Credit Suisse integration. UBS did not link the decline to AI.
How many AI applications does UBS have in production?
UBS reported 608 AI use cases in production, up 81% from a year earlier, and more than 1,090 additional applications in development. The release did not identify how many production cases are client-facing.
Source: ThorstenMeyerAI.com
