Audit a real app’s UI: find 10 flaws, fix 3
Every app you use daily has design flaws its team has gone blind to. Audit one real app yourself, then race a multimodal AI on the same screens — document the 10 best issues with screenshots, and redesign the 3 worst.
Posted by The AI Internships
The work brief
- 01Pick a real app you use often (not a global giant’s flagship — a local/mid-size app has more to find).
- 02Audit systematically, screen by screen: typography consistency, spacing rhythm, icon style mixing, contrast, touch target sizes, confusing labels.
- 03Then run the race: give a multimodal AI your screenshots and ask for its own audit — keep its real finds, throw out its hallucinations, note what it missed that you caught.
- 04Document 10 real issues (yours plus the AI’s best, clearly judged), each with a screenshot and the design rule it breaks.
- 05Redesign the 3 worst in Figma/Canva, side-by-side with the originals.
What you’ll produce
Submission standard
Submit the app, the 10-issue audit with screenshot evidence, and your 3 before/after redesigns, plus your AI workflow. Number your evidence and before/after redesigns to match your audit issues, and add a public design view-link plus an editor screenshot showing your account name and date.
The app audited and why you picked it
Short answerRequiredThe 10 issues: each with where it is, what rule it breaks, severity
Written responseRequiredScreenshot evidence of your 3 worst issues (one labelled image)
Screenshots must be from the REAL app (status bar / clock visible), and each must be numbered to the matching issue in your audit doc — stock or AI-generated 'app' screens fail.
Image uploadRequiredBefore/after redesigns of the 3 worst issues (one labelled image)
Each before/after must map to a numbered issue in your audit doc, so the redesign, the evidence screenshot, and the written diagnosis tell one consistent story. Redesigns that do not reconcile with the audit fail.
Image uploadRequired
You’ll complete these inside your private workspace.
What you must submit as proof
This brief requires evidence an AI can’t fabricate.
Public Figma/Canva view-link to your 3 before/after redesigns
Figma 'Share > Anyone with the link' or Canva view-link. Must open to YOUR editable redesigns — a dead or private link fails.
Public linkRequiredScreenshot of your OWN Figma/Canva editor building a redesign — account name and today's date visible
Show the live editor with your username and system date, not just the exported before/after.
Image uploadRequired
Submissions without this evidence cannot be submitted.
Protect other people in your proof. Blur faces, names, phone numbers and email addresses before you upload, and refer to anyone you worked with by role or number ("Listener 1", "the stall owner"). Your proof is only ever used to check your work — it is never published, never appears on your certificate, and is never shown in your public portfolio.
How your work is evaluated
The passing benchmark is 70/100.
Audit rigor
29%The 10 issues are real, specific, and correctly diagnosed.
Fix quality
29%The redesigns solve the diagnosed problem without breaking the app’s style.
Severity judgement
14%The 3 chosen as worst genuinely matter most to users.
Real audit, cross-checked
29%Auto-fail if the audit reads as a generic critique, if the evidence screenshots are not from the real app, if the redesigns have no live editable file, or if evidence/redesigns do not reconcile with the numbered issues in the audit doc.
How we grade your AI usage
Using AI is the point — it’s the skill this certificate proves. You’ll answer three short questions about how you used it: what you asked, what was wrong with its first answer, and what you changed. Specific, honest answers score high. “I pasted the brief and submitted the answer” scores near zero.