All micro-internships
Micro-internship brief ≈11h across ~3 days — run the test and let results accruelow-codeIndia edition

Run a real A/B test on a real audience

Marketing runs on experiments, not opinions. Design an A/B test, draft both variants with AI, run it on a real audience — your followers, class group, or community — collect real numbers, and call a winner honestly.

AI-assisted MarketingCampaign StrategyAnalytical ReasoningBusiness Communication

Posted by The AI Internships

The work brief

  1. 01Form a hypothesis (e.g. "a question hook beats a bold-claim hook for my audience").
  2. 02Create two versions that differ in ONE thing only. Use AI to help draft both, then edit them so nothing but your variable changes.
  3. 03Run it for real: two Instagram stories with polls, two posts at similar times, or a WhatsApp/Google Form split. Get 20+ combined responses/views.
  4. 04Record the real numbers, screenshot the results, and interpret them — including whether the sample is big enough to trust.

What you’ll produce

5 deliverables

Submission standard

Submit your hypothesis, both variants, the real results with numbers, screenshot proof, and your verdict. Small honest numbers beat big invented ones. Also include your AI workflow — tools, best prompts, and what you changed from the AI’s output. State the exact platform, dates/times, and audience for the test, and make sure your screenshot shows your own account, the counts, and the date.

  • Your hypothesis and the ONE variable you changed

    Short answerRequired

  • Version A and Version B, exactly as shown to people

    Written responseRequired

  • The real numbers: reach/votes/clicks per variant, and how long it ran

    Report reach/votes/clicks per variant and run-time; the figures must match your screenshot exactly.

    Written responseRequired

  • Your verdict: winner, confidence, and what you’d test next

    Written responseRequired

  • Exact platform, the dates/times the test ran, and who your audience was

    Must match the timestamps and account shown in your results screenshot.

    Short answerRequired

You’ll complete these inside your private workspace.

What you must submit as proof

This brief requires evidence an AI can’t fabricate.

  • Screenshot of the real results (poll counts, insights, form responses)

    Must show YOUR own account/handle, the poll/insight/response counts, and the date — matching your reported results.

    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.

Test design

14%

One variable changed; versions otherwise comparable.

Real experiment

29%

The test genuinely ran on real people, with numbers matching the screenshot.

Honest interpretation

29%

The verdict respects sample size and doesn’t overclaim.

Verifiable real test

29%

The screenshot shows the student’s own account, the counts, and the date, and the numbers/timeline reconcile across results, proof, and context. If the screenshot lacks account/counts/date, or the three don’t reconcile, this criterion is a fail regardless of the verdict’s rigour.

How we grade your AI usage

30% of your score

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.