🔍 Read the full analysis: How AI Agents Fit Into Daily Work At UBS on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
UBS reported that employees were running more than 12,700 AI personal agents in its third-quarter results on October 28, 2026. The bank also reported 608 AI applications in production and more than 1,090 in development, but did not explain what the agents can access or do. UBS CEO Sergio Ermotti said the bank does not allow outside AI models to act for clients.
UBS reported on October 28 that employees were running more than 12,700 AI personal agents to support daily activities, making the bank’s internal use of agent-style tools a measurable part of its technology deployment. The third-quarter results also listed 608 AI applications in production and more than 1,090 in development, but did not explain what the personal agents can access, whether they can take actions, or how their use is overseen.
In its third-quarter results release, UBS said the 12,700-plus agents were being run by employees “to support them in their daily activities.” The same section reported more than 18,000 software developers using AI tools in software development cycles, which the bank said led to measurable productivity gains. UBS did not quantify those gains in the release.
The bank reported 608 live AI use cases, an 81% increase from a year earlier, and more than 1,090 additional applications in development. It also cited nine “large-scale, end-to-end transformational AI initiatives” and said its US financial-adviser tool STAAT had generated more than 30 million insights so far in 2026. These figures are UBS’s reported totals; they do not establish how frequently each tool is used or what results it produces.
The scale of deployment had grown since July. In its second-quarter results on July 29, UBS said all employees had the ability to create their own AI agents. The bank’s latest release reports more than 12,700 agents in use, but does not say how many employees run them or whether individual staff members can operate multiple agents. Reuters reported the figures on October 28, and PYMNTS reported them the following day.
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
How Agent Use Reaches Bank Work
The figures put a concrete scale on a question facing large employers: how to allow staff to use AI helpers while controlling access to sensitive information and consequential actions. At a bank, an assistant that prepares a document summary differs materially from one that can send instructions, change account details or approve a transaction. UBS has not said which capabilities its personal agents have, so the reported count alone cannot show how much work they perform or the level of risk they create.
The disclosure also gives other companies a benchmark for discussing internal agent use, but not a ready-made model for governance. UBS says employees can create agents and reports thousands in operation; it does not publish the rules, review process or permissions behind them. The central operational question is therefore not just how many agents exist, but who can use them, what data they can reach and what checks apply.
UBS’s workforce figures do not show that AI has replaced jobs. The results said internal staff fell by about 1,500 full-time equivalents during the quarter, to 97,593. UBS and Reuters connected the reduction to the Credit Suisse integration. UBS did not attribute the decline to AI, and the reported figures do not establish an AI-related employment effect.
Popular right now
Breville Bambino Espresso Machine, Brushed Stainless Steel (BES450BSS)
Bose QuietComfort Ultra Wireless Noise-Cancelling Headphones with Spatial Audio, Black
Vitamix Propel Series 750 Professional-Grade Blender, 4 Programs, Black
As an affiliate, we earn on qualifying purchases.
From Employee Access to Agent Counts
UBS’s July statement said every employee had the ability to create AI agents, describing the effort 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 described access to a capability; the October results added a reported count of agents being run. Those are different measures: the bank has not said how many employees actually use the tools or how often.
UBS also describes other internal AI services. Its technology materials refer to Red, an in-house assistant rolled out to around 100,000 employees, and an AI-powered human-resources assistant in Microsoft Teams for tasks such as booking holidays and finding policies. These examples offer a picture of employee-facing tools, but UBS did not say whether they are included in the 12,700-agent figure or define “personal agent” in the results release.
On October 27, a day before the results, CEO Sergio Ermotti drew a boundary around outside AI models acting for clients. Reuters reported his remarks at an event at ETH Zurich. UBS also pointed to its sixth-place position in the 2026 Evident AI Index, a ranking of 50 large banks, and said it was recognised for talent development and responsible AI. Those rankings are separate from the bank’s reported agent count and do not describe the controls governing each agent.
“Employees are running over 12,700 AI personal agents to support them in their daily activities.”
— UBS, in its third-quarter results release on October 28, 2026
What UBS Has Not Disclosed
The results release does not define a “personal agent” or identify the tools and models behind the reported total. It does not say what information the agents can access, whether they can take actions without human approval, how outputs are checked, or which teams and tasks use them. The number of agents also does not reveal how many employees use them: one employee could potentially run more than one, but UBS has not clarified the count.
The bank has not broken out how many of its 608 production use cases face clients, how they are risk-classified, or how many are agent-based. Nor has it published a measure connecting the 12,700 agents to time saved, service quality, costs or staffing. The reported developer productivity gains are not quantified.
Reuters reported Ermotti’s statement that outside models are not allowed to act for clients, but the remarks do not explain the safeguards for employee-created tools inside the bank. The source material also notes that Swiss regulator FINMA expects banks to maintain a complete inventory of AI applications with risk classes, while saying that inventory work has become harder as applications become broadly accessible. UBS’s public results do not describe its inventory or review arrangements in detail.
The Next Test Is Oversight
UBS’s next public results or technology disclosures could clarify whether the agent count changes, what tasks the tools perform and what controls govern their use. The bank has not announced a specific date for additional information on its personal agents. Until it does, the October figures remain a measure of reported deployment rather than a detailed account of business impact.
For readers tracking the issue, the most useful follow-up details would be the number of employees using agents, the systems and data those agents can access, whether they can execute actions, and how humans review their work. A breakdown of the 608 live use cases by function and risk would also help distinguish internal assistance from tools that affect clients or financial decisions.
UBS’s disclosure points to a broader test for businesses adopting employee-built AI: whether oversight can keep pace with use. The bank has reported the scale of its deployment, while the operating rules and measured outcomes remain largely undisclosed. Those details will determine what the agent count means in practice.
Key Questions
How many AI agents did UBS report?
UBS said in its October 28, 2026, third-quarter results that employees were running more than 12,700 AI personal agents to support daily activities.
What can UBS employees’ personal agents do?
UBS has not defined the term or listed the agents’ capabilities. The bank’s public materials describe employee AI assistants, but do not confirm which of those tools are counted among the 12,700 agents or whether they can take actions independently.
Does UBS allow AI models to act for clients?
CEO Sergio Ermotti said, as reported by Reuters after an October 27 event, that UBS does not allow outside models to enter the bank and act for clients, such as managing accounts or moving deposits. His statement does not spell out all internal agent controls.
Did UBS link its staff reduction to AI?
No. UBS reported that internal staff fell by about 1,500 full-time equivalents in the quarter, to 97,593. UBS and Reuters tied the decline to the Credit Suisse integration; UBS did not attribute it to AI.
What else did UBS report about AI?
UBS reported 608 AI applications in production, more than 1,090 in development and more than 18,000 software developers using AI tools. It said developer tools produced measurable productivity gains but did not quantify them.
Source: ThorstenMeyerAI.com
