Skip to content
UxerProof

Product

A usability assurance system, not a screenshot tool

UxerProof models the journey, not the page. An AI User follows a Persona through a Circuit against an Environment you have verified, and the Run produces Findings, Evidence and a Proof Score you can compare against a Baseline.

Run your first CircuitSee a sample report

The vocabulary

One term per concept, used identically in the marketing site, the product and the API.

Circuit
A complete test.
AI User
A simulated participant that performs the journey.
Persona
The behaviour definition an AI User follows.
Project
A website or application under test.
Environment
A target deployment of a Project.
Run
An individual execution of a Circuit.
Finding
A discovered problem.
Baseline
An approved comparison state.
Evidence
A supporting artefact captured during a Run.
Proof Score
The overall, fully decomposed result of a Run.

Role Handoff Map

Demo data

Most serious failures happen between people, not inside a single screen. Role Circuits pass the journey from requester to reviewer to approver and show exactly where the handoff loses context.

Role Handoff Map

3 roles · 2 handoffs in this Circuit

MayaRequester
  1. Create request

handoff → reviewer

TheoReviewer
  1. Review evidence

handoff → approver

JunoApprover
  1. Approve

Before / After Proof

Demo data

Approve a Baseline, ship a change, and prove the difference dimension by dimension — including the things that got worse.

Before / After Proof

Guest checkout — single item

Baseline

Baseline — Checkout v3

run_4602

71

Proof Score

Candidate

Candidate — build 1184

run_4821

78

Proof Score

Dimension comparison between baseline and candidate
DimensionBaselineCandidateChange
Completion1001000
Friction4861+13
Clarity5870+12
Accessibility7174+3
Consistency8085+5
Journey duration301s214s-87
Open findings63-3

Resolved

  • Postcode format now shown inline
  • Bag confirmation persists on mobile

Regressed

  • Delivery surcharge moved one step later in the flow

A Proof Score you can argue with

Five dimensions, each traceable to specific steps and evidence. No hidden weighting, no vanity number.

  1. 01

    Completion

    Did the AI User reach the intended outcome of the journey?

  2. 02

    Friction

    How much hesitation, backtracking and repeated effort occurred?

  3. 03

    Clarity

    Were labels, states and system responses understandable in context?

  4. 04

    Accessibility

    Keyboard reachability, focus order, naming and contrast along the path.

  5. 05

    Consistency

    Did patterns behave the same way across steps and roles?

Circuit modes

  • Public Circuit

    Unauthenticated journeys across pages anyone can reach.

  • Authenticated Circuit

    Planned

    Journeys behind sign-in using short-lived, scoped credentials.

  • Role Circuit

    Planned

    Multi-role journeys that hand off between requester, reviewer and approver.

  • Mobile Circuit

    Planned

    The same journey replayed at mobile viewports and input modes.

  • Accessibility Circuit

    Planned

    Keyboard-only, focus order, contrast and semantics checks along the journey.

  • Regression Circuit

    Planned

    A Circuit replayed against an approved Baseline to detect drift.