Practice interviews

Would you ship this AI-written product claim?

The campaign is ready. Its strongest claim goes beyond the study.

The question

At fictional Vale Consumer Products in London, you own a moisturiser launch. An AI assistant proposes “Repairs the skin barrier in 24 hours for all skin types.” The launch is on Friday.

The evidence
Study
A controlled study found improved hydration after 24 hours in 60 adults with dry skin. It did not measure barrier repair or test all skin types.
Team
Marketing has booked media. The science lead can review revised wording today. A broader trial would take weeks.
Risk
The same claim will appear in product feeds used by AI shopping assistants.

“What would you ship on Friday, and what would you hold?”

A strong answer, step by step

Open each step in turn. This is one strong answer, not the only one.

Step 1: Match each claim to its test

What would you do here? Open the move you’d make.

  • Publish now and add a footnote to the study.

    Tempting, but it misses something. A footnote does not make an untested outcome true.

  • Separate the measured claim from the unsupported claim.

    The strong move. A fluent claim can silently change the outcome and the tested population.

See how a strong candidate says it

“Hydration is measured; barrier repair and all skin types are not. I would remove the unsupported parts before approving copy.”

What a strong candidate could say

Why it works: A fluent claim can silently change the outcome and the tested population.

Step 2: Keep a supported launch option

What would you do here? Open the move you’d make.

  • Delay the entire launch until a barrier-repair study covers every skin type.

    Tempting, but it misses something. The current evidence may support a narrower hydration claim. Do not turn the unsupported stronger claim into a reason to discard every supported launch option.

  • Change the wording to supports skin barrier health and retain all skin types.

    Tempting, but it misses something. Softer language still implies an outcome and population that this study did not test.

  • Ask the science and claims owners to approve narrower hydration wording with the tested population and limits.

    The strong move. This preserves a supported launch path while leaving the stronger barrier-repair claim on hold.

See how a strong candidate says it

“Ask the science and claims owners to review narrower hydration wording, including the population and study limits.”

What a strong candidate could say

Why it works: The choice is not limited to publishing the original claim or cancelling everything.

Step 3: Correct every distribution surface

What would you do here? Open the move you’d make.

  • Update approved wording across packaging, the website and retailer feeds, with a recorded evidence version.

    The strong move. Each surface can repeat the claim. A shared approved version and clear handoff prevent a rejected claim from surviving elsewhere.

  • Correct the website first and let retailer feeds update in their normal cycle.

    Tempting, but it misses something. Those feeds can keep distributing the rejected wording, including to shopping assistants, after the website is fixed.

  • Send revised copy to the agency and assume downstream platforms will replace it.

    Tempting, but it misses something. A handoff alone does not establish which version is live. Include the distribution owners and confirm the approved wording reaches each surface.

See how a strong candidate says it

“Update packaging copy, the website and retailer product feeds together. Record the approved wording and evidence.”

What a strong candidate could say

Why it works: An old product feed can keep spreading the rejected claim.

Step 4: Define the next evidence need

What would you do here? Open the move you’d make.

  • Repeat the hydration study with more adults who have dry skin.

    Tempting, but it misses something. A larger sample of the same endpoint and population would still not establish barrier repair for the intended broader claim.

  • If barrier repair matters, commission evidence that tests that outcome in the intended population.

    The strong move. The next study should close the specific gap rather than merely make the existing hydration evidence larger.

  • Use low complaint rates and positive shopping-assistant responses as evidence for barrier repair.

    Tempting, but it misses something. Those are distribution or experience signals. They do not measure the claimed biological outcome.

See how a strong candidate says it

“If barrier repair matters to the strategy, fund a study that measures that outcome in the intended population.”

What a strong candidate could say

Why it works: The research plan should answer the missing question.

The answer, in short

Decision
Hold the barrier-repair claim. Launch only with wording approved against the hydration evidence.
Because
The current study does not establish the outcome or population in the proposed claim.
What would change it
A valid study that tests barrier repair in the intended population could support a stronger claim after review.

Common traps

  • Treating persuasive copy as evidence.
  • Correcting the website but leaving retailer feeds unchanged.
  • Using the deadline to bypass the claims owner.

Check your answer

  • Did I distinguish hydration from repair?
  • Did I retain the study’s population limit?
  • Did I correct all distribution surfaces?
  • Did I name the missing test?

Last reviewed Nov 11, 2026