ABAmine Bouhlal

// Case study · 2024 — present

Conversational AI for consumer finance

The consumer-finance arm of a major European retail group

Delivered as Solutions Architect at NTT DATA

2

modes: self-service + agent assist

100%

answers grounded in enterprise knowledge

Regulated

financial-services environment

CUSTOMERSELF-SERVICEconversational agentADVISORAGENT ASSISTgrounded draftsGEMINI ENTERPRISEgrounded retrievalKNOWLEDGE BASEpolicy · productsCORE SYSTEMSaccounts · cards
Two modes, one architecture: customer self-service and advisor assist, both grounded in the same enterprise knowledge and core systems.

The problem

A card-and-credit operation where customers ask high-stakes questions — balances, payments, disputes — and every answer must be grounded in policy and account systems, not model guesswork.

The system

Architected conversational agents on Gemini Enterprise: retrieval grounded in the institution's knowledge base, secure integration with core customer systems, and an agent-assist mode that drafts grounded answers for human advisors instead of replacing them where regulation demands a person.

The outcome

Conversational customer experience and cognitive assistance in a regulated financial environment — automation where it's safe, augmentation where it isn't.

Gemini EnterpriseGCPVertex AIPython
Next case → A smart-city platform for 275,000 residentsDiscuss a similar system