Orange: AI consulting

Context and business problem

Orange has a large structure, with many departments, each responsible for a different product, and offers that have to reach both B2C and B2B customers. Producing marketing assets such as banners and getting them approved across these departments was slow and manual.

Constraints

• Many departments and a multi-step approval flow.
• Two audiences, B2C and B2B, with different offers.

My role and key decisions

1. AI automation processes

I propose automation processes: for example banner automation: what guidelines it needs, how to set it up in-house, and how to connect it to the approval flow across Orange’s product departments.

Decision: design the automation around the existing approval flow instead of bypassing it.

Why: departments keep control and adoption is realistic.

2. Offer landing pages

I prototype offer landing pages and improve the UI.

3. User-journey research

User-journey research for B2C and B2B.

4. Mailings designed around UX

I changed the structure, segmentation, CTAs and layout of the mailings.

How it was validated

Open and click-through rates on the mailings. The 2× was measured in an A/B test.

Outcome and impact

• Mailings redesigned around UX: open and click-through rates doubled.
• Clearer offer segmentation and a smoother experience across B2C and B2B.

What I learned

• In a large organisation, AI automation succeeds or fails on the process around it: guidelines and approvals matter more than the tool.