udacity/cd15542-autonomous-marketing-exercises/sa-product-accuracy

Solstice Active — product accuracy

A claim is either approved, banned, or unverified — there is no fourth category called "sounds fine." Plausibility is never a reason to approve. This is the safety net that runs before copy ships: it checks every factual and comparative claim against the tracker, so nothing goes out that Solstice can't back up.

What this skill does, and its boundary with brand-voice

This skill answers one question about each claim: is it substantiated? — meaning it matches an approved row in the tracker, stays within that row's limits, and carries its required caveat.

It is the factual counterpart to [[sa-brand-voice]], which governs how copy is phrased. When copy is going out, run both. A line can be perfectly on-voice and still make an unsubstantiated claim, and vice versa.

The tracker is the only authority

At the start of every review, read references/sa-claims-tracker.md. That file is the source of truth for what is approved, banned, or unverified. It changes over time.

• Never approve a claim from memory, or from the examples in this file.
• Applies to every claim in the copy, including claims that came directly from a customer's own quoted words in an interview. A customer saying something enthusiastic and specific does not make it substantiated — the source of a claim doesn't change whether Solstice can prove it.

Process — check the copy step by step

1. Extract every claim. Read the copy and list each factual or comparative assertion separately — fabric performance, fit/sizing, competitor comparisons, recovery/health effects. Quote each one exactly. If a sentence bundles two claims, split them.

1. Match against the Approved list by meaning, not wording. Copy paraphrases; the tracker won't contain the exact sentence.

1. Check for banned wording first — this is the centerpiece check. Solstice has never run a substantiated head-to-head comparison against any named competitor. Any claim that names a competitor and compares Solstice favorably against them — "better than [Competitor]," "outperforms [Competitor]," "unlike [Competitor], this actually..." — is BANNED outright, regardless of how the customer phrased it, how enthusiastic it sounds, or how plausible the comparison seems. There is no version of a named-competitor comparison that is currently approvable.

1. Check for upgrades / overreach. An approved claim pushed past its stated limit becomes a violation — e.g., "moisture-wicking" upgraded to "keeps you completely dry" or "supports recovery" upgraded to "cures muscle soreness."

1. Check required caveats. Some approvals are conditional — see the tracker's Limits column.

1. Anything unmatched → UNVERIFIED (hard block). If a claim matches no approved row, it cannot ship, regardless of how reasonable it sounds. Emit the exact tracker row that would need to be added.

1. Confirm every ban carries its reason. Each BANNED finding must cite the reason on file from the tracker.

Verdict for each claim

Assign exactly one:

• APPROVED — matches an approved row, within limits, caveat present (or none required).
• APPROVED — WITH REQUIRED CAVEAT — approved only once the missing caveat is added; state which.
• NEEDS CAVEAT — same as above when the caveat is simply absent from the copy.
• UPGRADE/OVERREACH — based on an approved claim but pushed past its limit; state the limit.
• BANNED — matches a banned row (most commonly: a named-competitor comparison); cite the reason on file. Give the in-voice, non-comparative replacement.
• UNVERIFIED — cannot ship — on neither list; emit the tracker row that would need to be added.

Output format

Return a per-claim table:

Then a one-line gate: SHIPS (all claims approved) or DOES NOT SHIP — escalate, followed by the list of blockers.

Do not invent proof

Never fabricate a proof source, a test result, or a comparison to make a claim passable. An unknown specific is UNVERIFIED, not a guess. A competitor comparison is BANNED regardless of how it's sourced, since Solstice has no approved comparative-testing claim in the tracker at all.

Known gaps to watch (blind spots)

• Named-competitor comparisons show up unprompted in customer interviews far more than in written copy. A customer volunteering "it's way better than [Competitor]" as an offhand compliment is common — treat it exactly like any other banned claim, not as a special case because it's "just their opinion."
• Recovery/performance claims are time- and person-specific. A single customer's outcome ("cut my recovery time") is an anecdote, not a substantiated brand claim, unless the tracker has an approved row for it.
• Approved ≠ unconditional. Several approvals only hold with their caveat. A bare version of an approved claim is not automatically shippable.