🔍 Read the full analysis: How UBS Brings Personal AI Agents Into Staff Work on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
UBS reported on Oct. 28 that employees were running more than 12,700 AI personal agents for daily work, alongside 608 AI use cases in production. The bank has not disclosed what data the agents can access, what actions they can take or how their use is reviewed.
UBS said Oct. 28 that employees were running more than 12,700 AI personal agents to support daily activities, giving a measure of how widely staff-facing AI tools have spread inside a major regulated bank. The results also listed 608 AI use cases in production, but did not explain what the personal agents can access or do.
In its third-quarter results, UBS reported nine large-scale AI initiatives, 608 live AI use cases and more than 1,090 applications in development. The bank said the number of live use cases had risen 81% year over year. It also reported that more than 18,000 software developers were using AI tools in development work, describing the productivity gains as measurable without giving a specific figure.
The same results said the bank’s STAAT tool had generated more than 30 million insights so far in 2026 for its U.S. financial-adviser teams. These figures are UBS’s own disclosures. Reuters and PYMNTS separately reported the 12,700-agent figure, but their coverage did not add detail about how the agents work.
UBS had reported a wider availability of agent-building tools in July. In its second-quarter results, it said all employees could create their own AI agents. The October figure refers to agents employees were running; it does not say how many staff members use them, whether some staff run several, or how the bank counts an agent.
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
What 12,700 Agents Mean for Bank Work
The figure matters because it moves the discussion of workplace AI agents beyond pilots and product announcements to reported use at scale inside a bank subject to extensive oversight. UBS has not said these agents make financial decisions or act autonomously. The number alone cannot show how much work they perform or whether they improve outcomes, but it signals that employees are using agent-style tools as part of daily routines.
For staff, the practical impact depends on the tasks involved. An agent that summarizes documents or prepares a meeting brief supports an employee; one that can send communications, change records or approve transactions would raise different control questions. UBS has not published enough information to establish where its personal agents sit on that spectrum. The distinction matters for assessing productivity, accountability and operational risk.
For other businesses, UBS’s disclosure highlights a management challenge: tracking tools that employees can create themselves. The relevant questions are not only how many agents exist, but what systems and data they can reach, what actions they may take, and who checks their output. The scale figure makes those governance questions concrete without answering them.
Products Worth Considering
Samsung 32-inch Odyssey OLED G8 (G80SH) 4K QD-OLED Gaming Monitor
LG UltraGear 27GX704A-B 27-inch QHD OLED Gaming Monitor
Samsung 49-Inch Odyssey G93SC QD-OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
From Employee Access to Reported Use
UBS’s July statement said all employees had the ability to create AI agents, with the bank arguing that this could help staff spend more time on judgment, collaboration and client service. The October results then reported more than 12,700 agents in use. The bank also describes an in-house assistant called Red, available to around 100,000 employees, and an AI-powered human-resources assistant for tasks such as finding policies and booking leave. The results do not say whether these tools are counted among the 12,700 personal agents.
UBS’s reporting showed other changes in its AI pipeline between the second and third quarters: live use cases rose from 560 to 608, while applications in development went from more than 920 to more than 1,090. The bank’s release does not provide a breakdown by business area, risk level or customer-facing status. Those counts describe different categories and do not, by themselves, show how many projects have reached employees or clients.
On Oct. 27, a day before the results, CEO Sergio Ermotti drew a boundary around external models acting for clients. Reuters reported that he said UBS was not allowing models to enter the bank and act on behalf of clients, such as managing accounts or moving deposits. That statement concerns outside models acting for clients; it does not define the permissions or safeguards for employee-run personal agents.
“Employees are running over 12,700 AI personal agents to support them in their daily activities.”
— UBS, in its third-quarter results released Oct. 28
What UBS Has Not Disclosed
UBS does not define “personal agent” in the results or specify which models and tools power them. It has not disclosed what internal systems or sensitive data the agents can access, whether they can take actions independently, or whether they only prepare material for a staff member to review. Without those details, it is not possible to tell how much authority the agents have.
The bank also does not say how many employees run agents, what proportion of the 608 production use cases are customer-facing, or how the agents are tested and monitored. Nor did it quantify the productivity gains it attributed to developers using AI tools. The headline count is therefore a measure of reported deployment, not a published account of economic impact or risk.
Swiss regulator FINMA expects banks to maintain an inventory of AI applications and classify them by risk. The source material says the regulator has noted that keeping such inventories has become harder as applications become more widely accessible. It does not establish whether UBS’s reported agents have caused a compliance problem or whether the bank has failed to meet those expectations.
The Next Test Is Agent Oversight
UBS has not announced a specific next milestone for its personal agents. The next useful information would be a clearer account of their permitted tasks, access controls, human review and risk classification, as well as a definition of how the bank counts agents. Further quarterly reporting could show whether the number continues to grow, but a count alone would still leave those questions open.
Investors, employees and regulators will also need to distinguish staff-facing assistance from systems that act on clients’ accounts or make decisions. Ermotti’s comments set out a stated restriction on external models acting for clients, while the third-quarter release reports employee use. UBS has not explained in these disclosures how its internal agent controls work or how that boundary is enforced.
Until the bank provides more detail, the confirmed development is a large reported staff deployment—not evidence that agents are replacing workers, independently handling client assets or producing a quantified financial return. Those outcomes remain unestablished by the published figures.
Key Questions
What did UBS report about AI agents?
UBS said in its Oct. 28, 2026, third-quarter results that employees were running more than 12,700 AI personal agents to support daily activities.
What can the personal agents do?
The results do not define their tasks or permissions. UBS has not said whether they only prepare or summarize work for staff, or whether any can take actions in internal systems.
Are the agents acting on clients’ accounts?
UBS did not say that its employee-run personal agents act on client accounts. CEO Sergio Ermotti separately said, as reported by Reuters, that the bank was not allowing outside models to act for clients, such as managing accounts or moving deposits.
Does the 12,700 figure show AI is reducing UBS staff?
No. The reported figure is a count of agents, not a measure of jobs removed. UBS’s internal staff count fell by about 1,500 full-time equivalents during the quarter, and UBS and Reuters linked the change to the Credit Suisse integration, not to AI.
How many AI projects does UBS have in production?
UBS reported 608 AI use cases in production in its third-quarter results, up 81% year over year. It also listed more than 1,090 applications in development, a separate category.
Source: ThorstenMeyerAI.com
