Testers with traits that matter.
Tech literacy, patience, device, accessibility, language. Identical testers tell you nothing.
- Describe, don’t configure
- One sentence in, and every trait and the background are drafted for you to check.
- Nothing saved unreviewed
- You review every line before a tester exists, and a lint pass flags a person who could not exist.
- Traits, not a script
- Patience, tech confidence and goals change how a tester behaves rather than what they say.
- Kept for the next study
- A tester you save can attempt any study in the workspace.
What a tester is made of
Pick a capability to see it drawn from the sample study.
Three testers ship with every workspace. The rest are yours, and each says how often it has run.
Who to send at a prototype first
The testers teams reach for.
Test with the people the room forgets
Low vision, a screen reader, large text, a right-to-left language, a slow connection. Pick them from the library and run.
“The label reads as cancel to me. I would not press it.”
Describe the user you already know
The cautious first-timer, the power user who compares three competitors, the professional in a hurry. One sentence each.
Find out which traits are worth recruiting for
Run the spread first. Where two testers diverge is where a human session is worth the scheduling.
A spread, not copies of you
- 6traits, all of them behavioural
Traits, not costumes
Every field on the form changes what happens in the session, or it would not be on the form.
- No contradictions found
Linted before it runs
A tester who is both a screen-reader user and a speed-reader is flagged rather than quietly simulated.
- Patience: lowPatience: high
Duplicate and adjust
Change one trait, run again, and the difference in the results is attributable to that trait.
Six fields, and every one shows up in the transcript
This is the whole form.
Tech literacy
How comfortable they are with software — the single trait that changes a session most.
Domain familiarity
How much they already know about your field, which decides whether your words help or hide.
Patience
How long they will keep trying before they give up and leave.
Primary device
Phone, tablet or desktop — what they are actually holding while they use it.
Accessibility needs
Low vision, a screen reader, large text, limited movement or a cognitive difference — the testers a team rarely gets to recruit.
Language and connectivity
Primary language, English comfort, right-to-left, slow connection.
About the testers themselves
Can I write my own AI testers?
Yes. Pick the traits, add a background, or describe the person in a sentence and let AI draft it. No raw instruction scripts.
How do I know the traits do anything?
Compare two sessions from one run. The spread is produced by the traits, not by chance.
What stops an AI tester contradicting itself?
A lint pass flags it before it runs.
Test from your pocket
The uTestMe tester app runs a study on a phone. Taps and screens recorded, consent first, results in the same view.
- Tap tracking
- Screen recording
- Consent first