Artefact builds generative AI customer-insight tool for a major French bank in one month

Overview

Artefact, a French AI and data consultancy, used IBM watsonx.ai to build a generative AI tool for a major French bank that lets professionals interact with AI-generated customer personas representing clusters of 2 to 3 million real customers, rather than working with anonymized raw data. Users define a target customer segment and engage the resulting avatar in a chat-like exchange to learn about consumption habits and preferences, with every AI-generated response made explainable. Artefact deployed the entire solution, from back end to front end, in just one month with a small team, and the bank is now pursuing a second generative AI project that would let the AI directly generate emails.

The challenge

A major French bank wanted to expand its service offering by harnessing customer data more effectively, but sharing entirely anonymized raw data with professionals had limited value for understanding consumer habits.

The solution

Artefact used IBM watsonx.ai to build a solution where users define a target customer segment and engage an AI-modeled persona avatar in a chat-like exchange to learn about consumption habits, with every AI-generated response made explainable and anonymized client data used for development.

Reported business value

Artefact deployed the entire solution, from back end to front end, in just one month with a small team; each cluster represents 2 to 3 million real customers, and the use case's success has generated interest and a second generative AI project featuring AI that directly generates emails.