Kering: a unified Power Platform layer across twelve brands.
A retail intelligence layer in Microsoft 365, used daily across twelve maisons. Operational data, automation, and brand-aware copilots — all inside their existing tenant.
The challenge
Twelve maisons. Twelve brand identities. One holding company that needed a single operational data layer without flattening what makes each house different.
Inside an enterprise that had already invested heavily in Microsoft 365, the brief was clear: extend what we own, don't rebuild it. Power Platform, M365, Power Automate, and the Copilot connectors were the canvas.
The harder constraint was governance — each brand controls its own data, its own tone, its own workflows. The shared layer had to be useful without crossing the line into central command.
Our approach
We started with a 4-week diagnostic across three maisons in parallel — different categories, different markets, different sophistication. The patterns we found there shaped the architecture for the remaining nine.
The architecture rule we set with the steering committee: shared infrastructure, brand-specific surfaces. One data layer, twelve UIs, twelve tones of voice.
The stack
- PlatformMicrosoft Power Platform — Dataverse for shared models, Power Automate for cross-brand flows.
- AI layerCopilot Studio + custom connectors over Claude for brand-aware generation.
- HostingExisting M365 tenant. No new vendor. No new procurement cycle.
- GovernancePer-brand RBAC, audit trails on every cross-brand read, opt-in joining.
Pilot to twelve brands
We piloted with three maisons over 8 weeks — one in fashion, one in jewelry, one in beauty. The first measurable wins were inside ops: weekly reporting cycles compressed from days to hours, retail floor analysis surfaced in Teams instead of buried in Power BI.
After pilot sign-off, the rollout to the remaining nine brands took 12 weeks. Each brand got a 2-week onboarding sprint. By week 24, all twelve were on the same shared layer.
What's next
Phase 2 expands the shared layer with category-specific copilots — clienteling for fashion, atelier scheduling for jewelry, sample tracking for beauty. The architecture is ready; each new copilot is a configuration, not a build.
“Yara Intelligence understood what most consultancies miss about luxury — that the operational efficiency conversation can't override the brand conversation. They built shared infrastructure that respected brand sovereignty.”
Tell us what you're trying to ship.
Book a 30-min strategy call. We'll tell you in 30 minutes whether AI makes sense for your business — and if it does, exactly what we'd build, in what order, for what budget. No deck. No follow-up email sequence. Just a real conversation.