UBS deploys AI programs to unlock efficiency
UBS is accelerating its AI investments and standardizing on a one-bank model as it continues integrating Credit Suisse through 2026. CEO Sergio Ermotti told investors the firm is rolling out large-scale AI programs to improve client experience, boost productivity, and strengthen operational resilience.
UBS launched more than 300 AI use cases across the firm in 2025. The bank also appointed its first chief AI officer, Daniele Magazzeni, to drive adoption and scale next-generation tools across front- and back-office processes.
Ermotti acknowledged a critical dependency: the final wave of client migration tied to the Credit Suisse integration. It carries the highest complexity and must land cleanly for UBS to fully retire legacy infrastructure by year-end 2026.
What this means for operations leaders
This is not a lab experiment. UBS is aiming for end-to-end process redesign, not isolated pilots. The focus is on standard platforms, common controls, and measurable gains in service quality and cost-to-serve.
Expect more automation of high-volume workflows, better triage for exceptions, and tighter control frameworks for model governance. With AI capabilities embedded into core processes, adoption will hinge on data readiness, change management, and line-level training.
Leadership moves to watch
There's movement across the industry as banks reset leadership around AI. Ronald Jansen, formerly global head of the AI lab in UBS's global banking unit, joined JPMorgan Chase to lead AI for HR and employee experience. Mike Dargan, previously UBS's chief technology and operations officer, left to become CEO of N26.
Other banks are formalizing AI roles as well. Wells Fargo named Saul Van Beurden head of AI and co-CEO of consumer banking and lending, and hired Faraz Shafiq to lead AI products and services. The Commonwealth Bank of Australia appointed Ranil Boteju as chief AI officer.
Practical actions for your team
- Lock in a one-bank operating model: standardize processes, controls, and platforms to limit variance and speed up rollout.
- Prioritize use cases by volume and waste: start with onboarding, KYC refresh, L1 service, reconciliations, and claims/exception handling.
- Fix data first: improve data quality, lineage, and access controls; establish clear rules for PII handling and model inputs/outputs.
- Stand up AI governance: model inventory, approvals, human-in-the-loop checkpoints, drift monitoring, and audit-ready documentation.
- Redesign end-to-end, not just steps: measure cycle time, straight-through-processing rate, error rate, and cost-to-serve for each process.
- Plan for migration waves: build runbooks, backout plans, customer comms, and dual-run economics well before cutover windows.
- Upskill the line: train frontline, middle office, and supervisors on prompts, tool hygiene, and exceptions handling with AI assistance.
- Track outcomes, not usage: tie AI metrics to productivity per FTE, client satisfaction, control findings, and infrastructure cost.
Risks and dependencies to manage
Leadership churn can slow decision-making, so clarify ownership early and define escalation paths. Migration complexity is a real constraint; sequence cutovers to protect client experience and control integrity.
Watch for model risk, privacy exposure, and vendor lock-in. Keep delivery in tight 90-day increments with clear acceptance criteria and business sign-off.
Why this matters
UBS is blending scale, governance, and process rework-exactly what it takes to turn AI into measurable operational gains. The emphasis on a single operating model and end-to-end redesign is a blueprint many operations teams can adopt.
For official updates and statements, see the UBS newsroom: UBS Media.
If you're building team capability, this curated list of AI tools for finance can help accelerate your short list and training plan.
Your membership also unlocks:
