PLATFORMS
AI personas
Simulate how your audience may react to a message or a scenario, and anticipate risks and opportunities before going public.
The problem
Challenge
Pretesting with real people is valuable, but it can be too costly for some teams, which then rely on judgement, and judgement is shaped by the people inside the organisation, who are rarely the audience. Most communication is therefore written, approved and published before anyone outside the room has seen it, and the audience’s reaction only arrives afterwards, when the budget is spent and the message can no longer change.
What it changes
Impact
Our AI Persona application has been tested and refined by our team of communications practitioners, data analysts and developers before being put to scale for our partners. It solely draws on existing data, interviews, field reports and social listening. Each persona carries rules about what it trusts, what it will not accept and what it has no way of knowing, so it does not answer with knowledge the real segment would not have.
Under the hood
How it works
From what you know about your audience to a message tested before it is published
01
Build the persona library
Each persona is defined from existing evidence about a real segment, with its context, beliefs, trust rules and sensitivities recorded alongside the sources it came from.
02
Set the simulation
You bring the draft messages, the scenarios, the hypotheses, whether that is three versions of an announcement, a set of headlines or policy responses, and select which personas should debate with another.
03
Run the conversation
The personas respond in role, saying what they understood, what concerns them and what would convince them. They can be questioned individually or run as a panel where they react to each other.
04
Assess responsibly
Results come back as transcripts, scores and a summary by an AI moderator outlining points of convergence and divergence between personas. This simulation can assist you in challenging your preconceived ideas.