Query digital twins built on real human personality — about anything.
P.H.I.L. is an endless lab full of synthetic people (P-twins, psychometric twins) built from real personality data ground truths. Put a question to them (e.g., your pricing, your pitch, your product, your name, your ad) and see how each kind of mind responds, and most importantly...why.
Write two versions. Pick an audience. Watch them react.
Demonstration model. Live PHIL scores each variant against validated 29-dimension profiles; this bench uses a simplified in-browser approximation so you can feel the loop.
Three steps between a hunch and an answer.
Built from real people
Every P-twin is grown from a real, consented psychometric profile — not a demographic guess. The panel thinks the way validated personality says people actually think.
Run your hypothesis
Drop in two versions of anything — a subject line, a headline, a value prop, a price — and PHIL scores each against every twin in your panel, thousands at a time.
Read the reasons
Not just a winner: the split by personality, the verbatim reasons each segment leaned the way it did, and the one change most likely to move the holdouts.
What's a P-twin?
A P-twin (i.e., digital twin with a personality) is a synthetic person built from a real, validated personality profile — a psychometric twin. It carries a full configuration of how someone weighs risk, detail, pace, and people, and it reacts to what you put in front of it the way that configuration predicts.
Run a thousand of them and you get what a focus group never could: scale and specificity at once. Not "women 25–34 in the Midwest," but "people who need proof before they move."
Every P-twin traces to a consented, de-identified profile. Panels are aggregated; PHIL never speaks for a named individual.
Speed, scale, and the why — pick three.
| Focus group | Demographic audience sim | PHIL | |
|---|---|---|---|
| Turnaround | Weeks | Hours | Seconds |
| Sample size | 8–12 people | Thousands | Thousands |
| Grounded in | A few real opinions | Demographic patterns | Validated personality |
| Tells you who | Sometimes | Yes | Yes |
| Tells you why | If you probe | Rarely | Every run |
| Tells you what to change | No | No | Yes |
Stop guessing which version lands. Ask the panel.
PHIL is opening to a first cohort of teams who'd rather test a hypothesis than argue about it. Put your name in.