Article published on January 30, 2025, in Choiseul Magazine - Joffrey Martinez, Partner & Global Financial Services Lead.
Interest rate cuts in Europe, geopolitical uncertainties, and the rising cost of credit risk… European retail banks are facing major challenges.
European retail banks are seeking new levers of efficiency to cope with the growing cost pressures. In this context, AI, particularly generative AI, offers concrete opportunities for time savings and cost reduction. However, truly realizing these promises of gains requires overcoming several challenges.
Choosing (and sometimes giving up).
The use cases of AI in banks are countless: process automation for compliance, fraud models, credit file analysis, and more. The challenge lies not in the proliferation of these opportunities, but in their selection. Each use case must be evaluated based on its potential for gain. An AI roadmap for the next 2-3 years, aligned with the overall strategy and revised every 6 months, along with a system for tracking value during execution, are essential for prioritizing and managing high-value projects.
Transform to win.
AI cannot simply overlay solutions onto existing processes. To achieve significant ROI, AI must be deeply integrated into processes by completely reshaping them and redefining tasks and roles. This transformation requires a multi-agent approach to orchestrate the completion of entire tasks rather than generating a sum of diffuse gains.
Measurement and control are key.
The number of AI/GenAI use cases in production within banks is set to explode. To keep IT costs under control, it is important to empower teams to take on part of the application maintenance for these solutions. The growing autonomy of AI solutions, particularly agents, requires enhanced governance to ensure both trust and efficiency, based on three pillars:
- Observation: IT services analyze interactions to detect issues.
- Evaluation: Refine models based on feedback from data scientists.
- Supervision: Identify recurring problems and implement fixes.
Here, measurement and control play a crucial role, surpassing even that of the model itself.
Involvement and support.
Training teams and adapting their skills are essential. The smooth adoption of these technologies will ultimately depend on their acceptance by the workforce and their alignment with business objectives.
AI provides banks with a powerful lever to tackle current challenges and turn constraints into opportunities. However, it is essential to activate the right levers to combine innovation and efficiency
By Joffrey Martinez
