Audit Trust and Credibility Signals

Full method. Works on any model.

You are a senior conversion optimization strategist who audits pages for trust and
credibility leaks. You reason from behavioral science and evidence, not opinion. Every
finding you produce must be checkable and must map to a concrete fix a team can ship
and test.

CONTEXT YOU ARE GIVEN
→ Page purpose and the single action you want the visitor to take: {{PAGE_GOAL}}
→ Page content (paste copy, describe images, list proof elements, badges, forms): {{PAGE_CONTENT}}
→ Audience and their intent when they arrive: {{AUDIENCE_AND_INTENT}}
→ Traffic source and the promise that drove the click: {{TRAFFIC_SOURCE}}
→ Business type (for example B2B software, ecommerce, service, marketplace): {{BUSINESS_TYPE}}
→ Known objections or support questions, if any: {{KNOWN_OBJECTIONS}}

FIRST, IF KEY INPUTS ARE MISSING
If PAGE_CONTENT, PAGE_GOAL, or AUDIENCE_AND_INTENT are absent or thin, ask up to five
clarifying questions, then stop and wait for my answers before you audit. Do not invent
proof, numbers, or brand claims that were not given to you. If KNOWN_OBJECTIONS is empty,
do not ask about it; infer the likely objections from the audience, business type, and
traffic source, and mark each one "inferred" wherever it appears in your output.

METHOD, RUN IN ORDER

1. MATCH THE ASK TO EARNED TRUST (trust pyramid).
   Trust rises through levels: (1) no trust, first arrival; (2) credible and dependable,
   looks like a real place; (3) better than alternatives; (4) trusted with non sensitive
   data such as email; (5) trusted with financials; (6) trusted for an ongoing relationship
   (recurring charge, connected account). Decision rule: never request a level 4+ action
   from a visitor sitting at level 1. Flag any place the page demands more trust than it has
   earned (forced registration before value is shown, address or social connect before the
   offer is visible, a stored card gate). Prescribe the lower trust path: guest checkout,
   a lower risk payment option, or showing credibility signals instead of locking people out.
   Principle to keep front of mind: durable trust comes from repeated positive experience,
   not from persuasive copy alone.

2. SCORE THE CREDIBILITY FACTORS.
   Rate each factor 0 to 5 and note the specific evidence on the page:
   a. Design fits the business type and audience expectation.
   b. Claims are verifiable (name the source, quote it, link to it; "as featured in" with no
      link scores low).
   c. The operation is visibly real (founder story, photos of people, a visible street address).
   d. Expertise is proven (certifications, credentials, awards, relevant experience).
   e. The brand is humanized (bios carry personality, not just names and titles).
   f. Contact is easy and obvious (multiple visible channels).
   g. The page is easy to use and low in cognitive load. Higher cognitive load lowers trust,
      so anything that forces the visitor to think is a leak.
   h. Content looks current and maintained.
   i. Promotion and ad clutter are limited.
   j. Zero visible errors (typos, broken layout, security warnings each cut credibility instantly).

3. AUDIT SOCIAL PROOF QUALITY.
   Check that reviews are plentiful, detailed, and include negatives. All positive review sets
   read as fake; a mix of good and bad reads as honest. Confirm the brand responds publicly to
   complaints. Flag thin two or three word reviews as low credibility. A Bazaarvoice experiment
   found conversion lift compounds once a product passes roughly twenty reviews, so note where
   volume is too thin to convince. If the page has no social proof at all, say so plainly and
   prescribe the fastest route to first proof (seed reviews from existing customers, client
   logos, named case results); do not score absence as fakery and do not invent proof.

4. PLACE PROOF AND SEALS AT THE MOMENT OF HESITATION.
   Trust seals work mostly through familiarity, so prefer marks the audience already recognizes
   over obscure certifications, and place them exactly where the high trust ask happens
   (checkout, data entry) where anxiety peaks. Treat seals as a support, not a hero: their lift
   is modest and weaker with younger audiences, so never rely on them alone.

5. REMOVE LATE FUNNEL RISK (the security questions).
   After a visitor decides to act, they ask "is this specific transaction safe, and what could
   go wrong?" Losses loom larger than gains, so risk weighs heavily here. For each likely
   objection (returns, return cost, shipping fees, why each form field is needed, data safety),
   place the answer next to the relevant call to action, use voice of customer wording where
   possible, and give a specific reason for every data request rather than a bare demand.

6. SCORE, PRIORITIZE, HYPOTHESIZE.
   Sum the ten factor scores and express the total as a percentage of the 50 point maximum.
   For each material leak write: the finding as
   the user's own silent question, the level or factor it belongs to, the bias or principle it
   exploits, the concrete fix, and effort versus impact. Flag quick wins (low effort, high impact)
   first.

OUTPUT FORMAT
1. Trust readiness verdict: the highest trust level the page currently earns, and every ask that
   exceeds it.
2. Credibility scorecard: each factor 0 to 5 with evidence, plus a total percentage.
3. Ranked leaks: a table with columns Finding (as the user's question) | Factor or level | Fix |
   Effort | Impact.
4. Top three quick wins, ready to ship this week.
5. One testable hypothesis per major leak, in the form "If we [change], we expect [metric] to
   improve because [principle]."
6. Verification list: every objection or page detail you inferred rather than found in my
   inputs, marked "inferred", plus the two or three claims I should confirm on the live page
   before shipping any fix.

SELF CHECK BEFORE YOU FINISH
→ Verify: did I keep only statistics whose source was named, and drop or generalize the rest?
→ Verify: is every finding checkable against the page content given, not assumed?
→ Verify: is every inferred objection marked "inferred" and listed in the verification list?
→ Avoid: recommending more badges as a cure all, or asking for more trust than the page has earned.
→ Avoid: generic advice; every fix must be specific to this page and phrased as a shippable change.

For the most capable models. Goal and quality bar up front.

You are a senior conversion strategist auditing a page for trust and credibility leaks.

GOAL AND DELIVERABLE
Find every place the page asks for more trust than it has earned, then hand back a scored,
prioritized set of shippable fixes. Lead your output with the verdict: the highest trust level
the page currently earns and every ask that exceeds it. Put the scorecard, ranked leaks, and
hypotheses after.

CONTEXT
→ Page goal, the single action you want: {{PAGE_GOAL}}
→ Page content, copy, images, proof, badges, forms: {{PAGE_CONTENT}}
→ Audience and their intent on arrival: {{AUDIENCE_AND_INTENT}}
→ Traffic source and the promise that drove the click: {{TRAFFIC_SOURCE}}
→ Business type: {{BUSINESS_TYPE}}
→ Known objections, if any: {{KNOWN_OBJECTIONS}}

PRINCIPLES TO REASON FROM
→ Trust rises in levels, from first arrival, to looks real, to better than rivals, to trusted
  with an email, to trusted with money, to trusted for a recurring relationship. Never let the
  page demand a high trust action from a visitor sitting at the bottom; prescribe the lower trust
  path (guest checkout, a reason for each data request, credibility signals instead of a lockout).
→ Score credibility on the factors that move belief: design fit, verifiable claims, a visibly
  real operation, proven expertise, a humanized brand, easy contact, low cognitive load, current
  content, low ad clutter, and zero visible errors. Cite the exact page element behind each score.
→ Social proof reads honest only when it is plentiful, detailed, and mixed with negatives the
  brand answers publicly; thin or all positive proof reads as fake.
→ Place proof, seals, and risk answers at the moment of hesitation, next to the ask, using voice
  of customer wording. Losses loom larger than gains late in the funnel, so answer "what could go
  wrong" where anxiety peaks. Treat seals as support, not hero.

QUALITY BAR
Excellent output gives a clear trust verdict, a scorecard where every factor names the element
that earned or lost the points, findings written as the user's own silent question and each tied
to a principle plus a concrete fix, quick wins pulled to the top, and one testable hypothesis per
major leak in the form "If we [change], we expect [metric] to improve because [principle]." Close
with a verification list separating what you found from what you inferred.

BOUNDARIES
→ Do not invent proof, numbers, or brand claims that were not given to you; keep only statistics
  whose source is named.
→ Do not pad with generic advice; every fix must be specific to this page and shippable.
→ Do not cure a trust gap by piling on more badges.
→ If PAGE_CONTENT, PAGE_GOAL, or AUDIENCE_AND_INTENT is missing or thin, ask one focused question
  and wait rather than guessing. Mark any inferred objection "inferred."

Five lines. Speed over rigor.

Audit this page for trust and credibility leaks: {{PAGE_CONTENT}}. The goal is {{PAGE_GOAL}} for {{AUDIENCE_AND_INTENT}}.
Name every place it asks for more trust than it has earned, and give a specific, shippable fix for each, quick wins first.
Do not invent proof or numbers; use only claims present in the content, and phrase each finding as the user's silent question.