Skip to main content
Methodology developed at Imperial College London

Extract data from your user testing sessions.

Testers run sessions independently. We capture real behavioural data. Data converts into actionable insights.

Safe. Secure. Scalable.

Start with a free scan. First conversion-friction findings in under 2 weeks.

Checkout completion

Illustrative
Neurotypical users87%
Neurodivergent users52%

A 35-point gap is the conversion you haven't seen, in the audience you haven't tested.

OpenScouter closes it in 2 weeks
Data controllerICO Registered
Social enterpriseSEUK Member
FoundersImperial College London
Data handlingUK GDPR

Everything you need to surface the friction your audit can't see

Live Session
Tester: ADHD Profile
Session recording
Emotion Voice Clicks

Frustration

High

Voice

Recording

Behaviour

3 rage clicks

  • 12 neurodivergent profiles matched to your product and user journeys
  • Dyslexia, ADHD, autism, dyspraxia, and 8 more conditions tested
  • Video evidence of real users navigating real tasks on your actual product
  • Not simulated. Not automated. Real people who experience barriers every day.
Learn more about Cognitive accessibility insights

See what your users actually experience, not just what they tell you

Traditional testing gives you what users say. Three data streams captured during every session reveal what they feel, how they move, and where they struggle.

See confusion before users report it

Users won’t tell you they’re confused — their face will. After every session, you’ll see exactly which moments caused frustration, confusion, or delight.

Emotion detection · 94% accuracy · 7 emotion classifications

Captures: Frustration spikes, confusion moments, delight signals

Find where users get stuck silently

Rage clicks, hesitation zones, and scroll abandonment reveal what surveys never capture. You’ll see where users get stuck — even when they don’t say a word.

Mouse tracking · Scroll depth · Rage click detection · Hesitation mapping

Captures: Navigation confusion, form abandonment, attention patterns

Hear what users actually say

What users say out loud while testing is gold. Every session is transcribed and tagged to specific WCAG criteria — so you know what to fix first.

Auto-transcription · Sentiment analysis · WCAG auto-tagging

Captures: Verbal frustration, think-aloud insights, accessibility barriers

Automated Accessibility Report

After every session, you'll get a prioritised list of what's actually blocking people, not a generic WCAG checklist. Severity-ranked, with fix recommendations and video timestamps.

Clone & Improve

We clone your site, then iteratively improve it

Your production site is never touched. We create an isolated copy, generate targeted CSS improvements with AI, and validate every change automatically before testers see it.

Isometric illustration showing a cloned website with three empty variant placeholders labeled A, B, and C
1

Your production site is never touched. Testing runs on an isolated copy.

2

Forms mocked, analytics stripped, third-party scripts removed. Testers never enter real data.

3

The clone is frozen at a specific moment, so every tester sees the exact same version.

Regeneration loop

When a tester says a variant did not help, AI uses their feedback to generate a better one. Up to 3 attempts per issue.

Tester rejects variant

"This didn't help with the date picker"

Feedback sent to AI

Rejection reason guides the next attempt

New variant generated

Re-validated through all 3 gates

Approved or skipped

Up to 3 attempts per opportunity

Built for organisations that need to show their working

Used by accessibility and inclusive-design leads in

Retail Banks
FinTech
Financial Services
Insurance
Comparison Sites
Healthcare
Government
Education

12 ND profiles assigned using stratified randomisation. Every test round covers the full spectrum of cognitive accessibility needs.

The deliverable

What our evidence looks like

Every finding ships with cohort impact, behavioural deltas, the WCAG criterion it touches, and a fix your engineers can plan against. A pilot delivers 30 to 50 of these per journey.

This is one finding from a recent pilot. Client and product details are fictionalised; format and severity model are identical to live engagements.

Read the full methodology
CriticalWCAG 3.3.2 · Labels or instructionsMortgage application

Form abandons spike when “gross income” is undefined for first-time applicants

On step 3 of the affordability calculator, four of six testers either dwelt for over a minute, reset the form, or fabricated a value. The field has no inline definition or worked example.

Cohort
4/6

struggled

Dwell
81s

baseline 23s

Abandon
47%

baseline 14%

Recommended fix

Inline disclosure under the field with a worked example (salary + bonus + dividend income). ~2 hours engineering, ~1 hour content.

The honest scope

Where we fit in your accessibility stack

Cognitive accessibility is one part of a complete accessibility programme. We are deliberately specialised in that part. The two columns below are how we describe the offer in proposals and on procurement calls.

What OpenScouter delivers

  • Cognitive friction in journeys: ambiguity, overload, sludge
  • Behavioural evidence per ND profile (ADHD, dyslexia, autism, dyspraxia)
  • Task-success and abandonment metrics with cohort splits
  • Mapping to the WCAG 2.2 success criteria cognitive issues touch
  • Iterative A/B retesting until measurable improvement
  • Methodology trail suitable for procurement audit

What you still need elsewhere

  • Screen-reader, magnifier, or switch-device testing
  • Keyboard-only navigation conformance
  • Full WCAG 2.2 AA conformance certification
  • EAA, Section 508, or EN 301 549 conformance reports
  • Authoring your formal accessibility statement
  • Verification of AT-dependent criteria (e.g. 4.1.2 Name/Role/Value)

<2wk

Report Turnaround

From briefing to findings

12

ND Profiles Per Round

Full-spectrum coverage

3x

More Issues Found

vs automated tools alone

£17.1bn

UK Click-Away Pound

Spend lost annually to inaccessible sites

Scouty mascot with encouraging expression

Ready to see what 1-in-5 of your customers experience?

Start with a free scan, or talk to us about a neurodivergent user testing pilot. First findings in 2 weeks.