BEBahae Eddine
back to projects
Enterprise AI · CRMConcept

Attijari AI Banker — Enterprise AI CRM

Concept case study with Attijariwafa bank (Banking & Financials) — an AI-powered CRM that scores clients, surfaces intent, and turns every relationship manager into a data-informed banker.

stack
PythonFastAPILangChainPostgreSQLAzure AI FoundryNext.js

Concept case study developed under XAI for Attijariwafa bank — Morocco's premier banking and financial group, with operations across Africa and Europe. The brief: relationship managers drown in data but starve for insight. The next best action is hidden in a dozen systems, and only the most experienced bankers can find it.

Problem

A bank's CRM holds the raw material — client profiles, transaction history, product usage, service interactions — but almost none of it is actionable in the moment. A relationship manager preparing for a client meeting has no single answer to "what does this client need today, and what should I offer?" The result is missed cross-sell, generic outreach, and value left on the table.

Approach

An enterprise AI CRM that turns banking data into a next-best-action engine:

  • Client 360: transaction history, product holdings, and interaction signals unified into a single, explainable profile.
  • AI scoring: models score each client on intent, propensity, and risk — so prioritization is driven by data, not memory.
  • Next-best-action recommendations: the system suggests the right product, channel, and timing for each contact, personalized to the client's life stage and behavior.
  • Conversational layer: an AI banker assistant that helps staff summarize a client file, prepare for a call, and draft compliant follow-up messages.
  • Compliance by construction: role-based access, audit trails, and boundaries that keep the assistant out of anything it shouldn't touch — banking is a regulated game, and the system plays by the rules.

Key decisions

  • Explainability is non-negotiable. A banker cannot act on a score they can't explain — every recommendation carries its reasons.
  • Human in the loop. The AI proposes, the banker disposes. Automation handles the busywork; judgment stays with people.
  • Owned data. The platform runs on the bank's infrastructure, keeping customer data inside the organization as regulation requires.

Status

Concept case study — representative of the enterprise AI/CRM engagements XAI delivers, published as part of the studio's portfolio.