The answer lives in this podcast

How did Ana-Maria Thijssen use Microsoft Copilot to develop audience personas for a global change communications programme?

21 August 2026 Unprompted: Real AI in IC · Ana-Maria Thijssen

← Voir la fiche Unprompted: Real AI in IC

Thijssen's team fed documents, org charts, communication models, and examples into Microsoft Copilot and asked it to identify the most distinct personas from that material — landing on around seven or eight groups covering functions such as finance, sales, and purchasing. They then used those personas to tailor communications to each group's specific needs, risks, and preferred channels, and trained local change coordinators to adapt global templates to their own audiences.

Feeding raw programme material into Copilot — and letting it find the patterns

Running change communications across a global SAP transformation that will ultimately reach 6,000 users means one thing above all: you cannot treat your audience as a single block. Thijssen's approach was to let the source material do the work. Rather than building personas from scratch through workshops or surveys, she fed the existing programme documents directly into Microsoft Copilot as described in this episode of Unprompted: Real AI in IC and asked it to surface the most meaningful distinctions.

The inputs were practical and already available: org charts, communication models used by the programme, and real examples of communications already produced. Copilot analysed that material and returned a set of distinct personas — roughly seven or eight in total — mapped to functional domains like finance, sales, and purchasing. No blank-page thinking required. No external consultants. Just structured prompting against documents the team already owned.

"AI became my virtual colleague. It's not a person that does my job. It's more like an entity that helps me bounce ideas and not feel so alone."

From personas to local action — a framework that travels across 30 locations

Identifying the personas was only step one. The real value came from what Thijssen did with them: each persona was mapped to its specific risks, information needs, and preferred communication channels. A finance audience has different concerns about an SAP rollout than a sales team or a purchasing function — and the communications they receive needed to reflect that, as Thijssen explains in full in the episode.

With a programme now expanding to 30 locations in the next wave alone, Thijssen could not personally manage every local communication. So the persona framework was built to be handed off. Local change coordinators were trained to take the global templates and apply them to their specific local audiences — using the persona structure as a guide rather than reinventing the wheel at each site.

This is what makes the Copilot-driven approach structurally sound: it produced an output that a single communications professional could create alone, document clearly, and transfer to others. The AI did not replace the strategic thinking — it compressed the time needed to get there. This kind of practical, scalable use of AI in change communications is exactly what Unprompted: Real AI in IC was built to surface.

It is also worth noting the scale of the challenge Thijssen was solving. Joining a programme mid-flight — over a year and a half after launch, with colleagues who had been embedded since January 2015 — meant she had no luxury of slow ramp-up. Using Microsoft Copilot to process and categorise existing programme material was not just efficient: it was the only realistic path to producing rigorous persona work at pace, without a full team behind her. The approach is covered in detail across S01 EP5 of Unprompted: Real AI in IC.