AI UX diagnostic

AI products lose users before they build trust.

Only 46% of people trust AI.1 Most features ship into that environment without accounting for it. Then teams watch activation and retention slide without knowing why.

Fully async 72-hr turnaround Fixed price No meetings
0 %

of people trust AI1

Most AI features ship straight into that gap.

Teams watch activation drop and retention slide without a diagnosis. The assumption is always the feature. It's almost never the feature. It's trust. Trust is a UX problem with specific, fixable causes.

The gap

What's actually happening

Activation drops. Retention slides. AI features go unused. The failure points are specific and repeatable.

The trust window is narrow.

Users form their opinion of an AI feature inside the first two sessions. After that, re-engagement is very hard to earn back.

  • AI outputs users can't verify or understand
  • Onboarding that doesn't establish credibility before asking for input
  • Error states that signal incompetence, not limitation
  • AI features shipped without ever testing the trust moment

Teams see the symptoms, not the cause.

Usage data shows the drop-off point. It doesn't tell you whether the problem is copy, interaction design, transparency, or something else entirely.

A prioritized diagnosis is the missing piece.

Not a generic checklist. A specific ranked list of what's breaking trust in your product, with a clear signal on where to start.

The process

Three steps, 72 hours.

01

Answer three questions.

The qualifier below tells you which diagnostic fits your product and situation. Takes about 90 seconds.

02

Book and send a link.

No calls scheduled. No kickoff meeting. You send a share link to your product and any context you want included. That's it.

03

Receive the report.

A prioritized findings document specific to your product. What's costing you users and where to start fixing it, in 72 hours.

Frank Rodriguez, Lightwell
The practitioner

One person does the work. You know exactly who reviewed your product.

Frank Rodriguez has spent years in AI product design, watching trust problems end promising features before they found their audience. Lightwell exists because the diagnostic work that actually fixes these problems doesn't happen anywhere else at this scope and speed.

Every diagnostic is run personally. There's no team behind it, no output delegation, no AI-generated findings. What you receive is a direct clinical read of your product's UX. Specific, prioritized, and ready to act on.

Frank Rodriguez

Diagnostics Practitioner · Lightwell

hello@lightwellhq.com
What you get

Three deliverables. Every diagnostic.

01

Prioritized findings report

Not a checklist. A ranked document of what's costing you users. Specific, actionable recommendations for each issue.

02

Annotated design file

Every screen reviewed with findings marked directly on the design. Shareable with your team the moment you receive it.

03

Recorded walkthrough

A narrated screen recording of the findings, run over the actual product. Watch it yourself or share it with anyone on the team.

Sample diagnostic

See what a walkthrough looks like.

lightwell-diagnostic-walkthrough.mp4

Sample walkthrough: The Budget App

Pricing

Three diagnostic tiers

Focused diagnostic
$595

Single AI flow · 72-hour turnaround

One specific AI feature or flow. For teams that know exactly where the problem is and need a fast, precise answer.

  • One AI feature or flow reviewed
  • Prioritized findings report
  • Annotated design file
  • Recorded walkthrough
Coming soon
Full funnel audit
$3,500

6-hour cap · 72-hour turnaround

Full product funnel, signup to retention. AI feature UX across the entire product. For teams with signal but not a diagnosis.

  • Full product funnel, signup to retention
  • AI feature UX across the product
  • Prioritized findings report
  • Annotated design file
  • Recorded walkthrough
Book the Audit
Find your fit

Not sure which tier fits?

Answer three questions. Takes about 90 seconds.

What best describes your product?
What do you most want to audit?
How many distinct AI features or flows do you want reviewed?

Please answer all three questions.

Your recommendation

What's included

Prioritized findings report
Annotated design file
Recorded walkthrough
72-hour turnaround

Ready to get started? Drop your details and we'll take you to booking.

After booking, complete a short intake form to share access and context.
Preview the intake form

1 KPMG and University of Melbourne, Trust in Artificial Intelligence: Global Insights 2025, 2025. 46% of respondents reported they trust AI systems.