Insights

Thinking on evidence-grounded AI
and how decisions actually get made

Practical writing on AI personas, qualitative research, decision support and the gap between knowledge and the choices teams make with it.

Featured Research & methodology · 8 min read

Why audience research gets filed after the brief, and what to do about it

Organisations invest substantially in qualitative and quantitative research. Within weeks of the brief, most of it is archived. The decisions that follow are made from memory and assumption. This piece explores why that happens and what a different approach looks like.

Read the article

5×

Average rework rounds
when research isn't active

Explainer · 10 min read

What are AI personas, and how do they differ from traditional user personas?

AI personas built from real research data behave fundamentally differently from the static documents most teams are familiar with. This explainer covers what they are, how they work and where the distinction matters.

Comparison · 9 min read

AI personas vs synthetic users: what's the difference and when does it matter?

Both promise to replace time-consuming human research. They take very different approaches, with very different implications for when you can trust what they produce. Here's how to tell them apart.

Technical · 7 min read

Why AI outputs need confidence scores, and what a good one actually tells you

Accuracy percentages are fine for benchmarks. In live professional use, what you need to know is how much of this specific response is grounded in evidence, and how to act when that figure is low.

More articles on AI research methodology, persona design and decision intelligence coming regularly.

See the platform in practice

The best way to understand how evidence-grounded AI works is to see it applied to a real use case and evidence base.

Book a demo Case studies