Case study · 01 / 11

Gagnon Frères

Internal software and AI for a retailer

Fig. 01 — Gagnon FrèresHover for colour

Digital innovation specialist in a five-person IT team: I design and build the internal software, applied AI and in-store tools of a Quebec retailer with 7 stores.

Since February 2026, I have been a digital innovation specialist at Gagnon Frères, a Quebec retailer with 7 stores, in a five-person IT team. My role: find where software and AI save time, then ship. I designed and built, end to end, the teams' internal platform, an in-store customer reception tool and a tablet sales app connected to the ERP. I am also building an AI assistant that answers plain-language questions from company data. Alongside that, I work in the Dynamics NAV ERP, handle the Azure architecture, the whole GitHub organization and repository security. These tools and their data are internal: this page stays deliberately general, with no screenshots or code, and the visuals are diagrams.

Main Features

Technical Challenges

Technologies

Year

In the wild

Diagrams · no screenshots of internal tools

What it does

01Internal platform for the teams: designed and built end to end

02Decision-support tools for buyers, with automatically refreshed data

03In-store customer reception managed in real time across 7 stores, on tablet, mobile and front desk

04Assisted selling on tablet: customers follow their file and sign on screen, final documents are produced automatically

05AI assistant that answers plain-language questions from ERP data

06Dynamics NAV ERP development and upgrades, point of sale included

07Azure architecture and technology choices, with cost assessment

08Full GitHub organization: move to GitHub Team, rules, roles, CI/CD and team guide

09Repository and secrets security, monitoring and observability

10AI pilot for the team (Copilot, agents) and best-practice training

How it was solved

01Making modern web tools talk to an ERP that has been running for years

02Shipping tools used every day in stores, on varied devices, without heavy training

03Putting AI to work on concrete questions without exposing sensitive data

04Setting GitHub and security rules in a small team without slowing it down

05Choosing technologies that stay affordable to run

06Keeping a polished design, even for tools only the team will see

Stack

Front-end

Back-end

Tools