Santander Opens AI Access To 185,000 Staff As Banking Automation Faces Its Scale Test
Santander is giving all 185,000 employees access to AI tools after reporting €35 million in first-quarter value, but wider automation in fraud, KYC, payments and voice channels still has to prove governance at scale.
Santander Opens AI Tools To Its Whole Workforce
Santander is extending AI access to all 185,000 employees, turning its bank-wide AI programme from a specialist rollout into a broader operating test for productivity, fraud handling and customer service.
The Spanish bank says the technology should generate more than €200 million in business value in 2026.
It has already attributed €35 million in first-quarter value to AI work, while keeping a wider target of more than €1 billion between 2026 and 2028.
The scale of the rollout is the main change.
About 40,000 Santander employees are active AI users today, but access is now being widened across the group.
Microsoft Copilot is part of the employee productivity stack, while the bank is also using a multi-provider model that includes OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini, startups, other technology partners and G42.
The Bank Is Measuring AI Through Workflows
Santander is not presenting the programme only as a staff-assistant deployment.
The bank says more than 280 process automation agents are already in production, covering manual work and end-to-end workflows in credit, fraud, KYC and operations.
Brazil shows where the bank is trying to attach AI to measurable operating outcomes.
For card fraud claims, Santander puts the speed gain at around 95%, says automation can reach up to 90%, and reports an error rate below 1%.
Those figures give the rollout a stronger evidence base than a simple software-access announcement.
They also show the risk for banks adopting AI at scale: productivity claims become more credible when they are tied to a named workflow, but they also invite closer scrutiny over accuracy, controls and customer impact.
Voice Channels Are The Next Customer-Service Test
Santander’s UK arm is starting to introduce AI into voice channels for card-related queries.
The bank’s target is around 240,000 calls, equal to 40% of annual volume, handled through self-service.
If the rollout works as planned, Santander expects customers to save around 26,000 hours and service teams to regain around 45,000 hours for more complex cases.
Those are operational claims, not proof that customers will prefer the automated route, and the source material does not give customer-satisfaction results for the voice-channel rollout.
Payments unit Getnet is another part of the programme.
Santander says AI is being used to support international customers and merchants, including cardholders paying abroad who may prefer to pay in their home currency.
The bank links that work to cross-border payment services, customer experience and merchant conversion, but does not give a revenue figure for Getnet from the AI use case.
The Open Question Is Governance At Scale
Ricardo Martín Manjón, Banco Santander’s chief data and AI officer, framed the programme as a move beyond theory, saying the bank can convert selected AI capabilities into value across the group.
For financial institutions, that scale is also the harder part.
Santander is moving from targeted AI use to access across the whole workforce, while deploying automation in fraud, KYC, operations, payments and customer service.
The unresolved issue is whether the bank can keep the measured gains visible as more employees and workflows join the programme.
Santander has given value targets and operating metrics, but the broader rollout will still depend on governance, error control and whether customers accept more automated banking interactions.
