SpainFinance and insurance2026 – ongoingSantander
Specialized AI models in experimental customer service
Santander reported more than 50% lower costs in an experimental AI customer-service use case
Problem
Santander wanted to reduce the cost of running AI models for banking tasks while preserving response quality.
Approach
Santander AI Lab worked with NVIDIA to evaluate models and specialize them for banking tasks. It built Model Arena for comparisons and used NVIDIA NeMo Customizer for specialization. In an experimental customer-service use case, the bank reported better conversation quality and lower costs with a more efficient model.
Results
Takeaway
The reported improvement came from an experimental use case. It did not establish equivalent savings across customer-service operations.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- Customer serviceIT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AIConversational and voice AI
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
Santander described model evaluation and specialization work with NVIDIA. It reported improved conversation quality and a cost reduction of more than 50% in an experimental customer-service use case.
Full details are in the original sources.
