Santander
Santander reports 35 million euros of AI value and 280 agents in production
Santander says its AI-first programme generated 35 million euros of business value in the first quarter of 2026. It reports more than 280 process automation agents in production and has opened AI access to all 185,000 employees.
Key figures
Copied exactly from the sources. Copy any figure with its citation.
€35 million
AI business value generated in the first quarter of 2026, as measured and reported by Santander1
More than 280
Process automation agents in production across credit, fraud, KYC and operations1
40%
Share of all code developed by AI in June 20261
over 17,000
People working with agentic AI in software development in May 20261
What it took
Cost, team size and timeline were not disclosed in the sources.
The problem
Santander wants AI to make the bank faster, safer and more efficient, and to bring in new growth. The bank's Chief data & AI officer describes the shift as moving from AI ambition to execution, with solutions that start in one country or function and are reused across the group.
What they built
The programme has three parts. First, process automation agents that handle manual tasks and end-to-end workflows in areas such as credit, fraud, KYC (checks on new customers) and operations. Second, AI in software development. Third, AI tools for everyday work: Microsoft Copilot for productivity, plus a multi-provider approach for more specialised uses.
How they did it
Santander set a target of more than 1 billion euros in business value from AI between 2026 and 2028, through added revenue and cost reductions. Its stated approach is to focus on fewer things, measure their impact and scale what works across the group. Until now nearly 40,000 employees actively used AI tools; from 22 June 2026 access extends to all 185,000. Training, practical guidance and learning communities support adoption. The bank says it does not share customer data externally to train third-party models.
Results
Santander reports 35 million euros of business value in the first quarter of 2026 and says it is on track to exceed 200 million euros by year-end, which is a forecast, not a result. It says more than 280 process automation agents are in production, over 17,000 people worked with agentic AI in software development in May 2026, and 40% of all code in June 2026 was developed by AI.
All reported figures
| Measure | Figure | Citation |
|---|---|---|
| AI business value generated in the first quarter of 2026, as measured and reported by Santander | €35 million1 | |
| Process automation agents in production across credit, fraud, KYC and operations | More than 2801 | |
| Share of all code developed by AI in June 2026 | 40%1 | |
| People working with agentic AI in software development in May 2026 | over 17,0001 | |
| Employees with AI access worldwide after the rollout | 185,0001 |
Lessons for your project
Define a value measure and report it every quarter. Santander counts added revenue and cost reductions and published a first-quarter figure.
Focus on fewer things that move the needle, measure them and scale what works. That is the bank's own stated logic.
Treat access as the start. Santander pairs tools with training, guidance and communities where staff share examples.
Start solutions in one country or function but design them for reuse across the group.
Caveats
- The only source is a post by Santander's Chief data & AI officer, so all figures are the bank's own and not independently audited.
- The post defines 'business value' as additional revenue plus cost reductions but gives no method. The more than 1 billion euro target for 2026 to 2028 and the expected 200 million euros by year-end are goals, not results.
- The post does not say which models power the process automation agents. It names Microsoft Copilot, OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and G42 as part of a multi-provider approach, not as the models behind a specific figure.
- The 35 million euro figure measures value from AI in general. The post does not say how much of it comes from generative AI or large language models, and it also mentions machine learning in other examples.
- The post does not say how many of the 280 agents use generative AI, so the agent count may include older automation.
- The announcement is dated June 2026 and extends access to all employees; the programme started earlier. The deployed date shown is the announcement month.
- No cost, team size or timeline is disclosed.
Sources
- Santander turns its AI-first strategy into measurable impact and extends AI access to all 185,000 employees
Santander · 2026-06-22 · Company blog · santander.com
Summarized in our own words from the sources above. Spotted an error? See corrections.