Turn product facts into value proposition candidates
Connect verified capabilities to audience outcomes without presenting assumptions as proof
8 min setupTested with:ChatGPTReviewed: 2026-08-28
1
Add your context
Your text stays in this browser. AILesson Prompts does not send it to a model or server.
2
Your prompt
Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt
Create value proposition candidates using only the supplied evidence.
Verified product facts, limits, price, availability, performance, and proof:
[facts]
Audience segment, jobs, problems, alternatives, outcomes, objections, samples, and unknowns:
[audience]
Channel, stage, tone, length, prohibited claims, legal rules, and decision criteria:
[use]
Build an evidence chain for each candidate: audience situation → problem or desired progress → relevant capability → plausible benefit → supplied proof → limitation. Label every link confirmed, inferred, or unknown. Do not convert a feature into guaranteed time, money, health, status, or performance gains unless direct evidence supports that exact claim and population. Produce meaningfully different positioning angles, not synonyms. For each include a concise headline, one-sentence explanation, proof line, qualifier, intended segment and context, likely objection, evidence strength, and claim risk. Rank by audience relevance, differentiation, evidence, clarity, and fit to the stated channel. Propose a falsifiable validation test and decision threshold for uncertain links. List excluded claims and why.
Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.
From input to outcome
A worked example
See how concrete context turns this recipe into a usable result
Actual input
Verified product facts
Meeting-notes app. Records with consent; transcribes English; lets users highlight a transcript and assign a follow-up item. Internal test: median draft transcript available 90 seconds after a 30-minute recording, n=40, stable Wi-Fi. No evidence of accuracy across accents. USD 9/user/month. Zoom only. No CRM integration.
Audience evidence
Six interviews with small remote agencies: five manually copy follow-ups after client calls; three reported missed actions; four worried automated notes may be inaccurate. All six use Zoom. Sample came from current newsletter. Desired outcome: leave calls with a reviewable follow-up list, not eliminate review.
Use and claim constraints
Homepage hero for US prospects. Plain, specific, maximum 12-word headline. No claims of perfect accuracy, no 'never miss anything,' no customer quotes, no guaranteed time savings. Rank for evidence and relevance.
Example output
Reviewable follow-up angle — strongest fit
Headline: “Turn Zoom conversations into a follow-up list you review”
Explanation: highlight transcript moments and assign follow-up items after an English-language Zoom call.
Proof and qualifier: in an internal stable-Wi-Fi test, the median draft transcript for 30-minute recordings was available within 90 seconds after the call; n=40. Human review remains necessary, and accent accuracy was not measured.
Evidence chain: Zoom use and manual follow-up are confirmed in six interviews; capability is verified; a more reviewable workflow is plausible but not yet behaviorally validated. Main objection: transcript accuracy. Claim risk: low to medium.
Fast draft angle
Headline: “A draft transcript shortly after your Zoom call”
This has strong internal timing evidence but weaker connection to the audience's actual job than candidate 1. Keep “draft,” the test context, and the absence of an accuracy claim.
Action-assignment angle
Headline: “Highlight the moment, assign the next action”
Capability is verified and concrete, but evidence that assignment itself matters to this segment is incomplete. Test before making it the lead.
Validation: randomly show qualified Zoom-using agency visitors candidates 1 and 2. Predeclare primary measure as the proportion starting a product walkthrough and a minimum sample; those are currently missing. Follow with five task sessions measuring whether users can create and correctly review a follow-up list. Do not claim success from clicks alone. Exclude “never miss an action,” “perfect notes,” “works with every accent,” “saves hours,” and CRM-sync claims because the supplied evidence does not support them.
Why this works
1
Explicit evidence chains expose the unsupported leap between capability and customer outcome.
2
Distinct angles and validation rules turn copy exploration into a reviewable decision.
Check the result
Can every factual phrase be traced to supplied product or audience evidence?
Are inferred benefits visibly qualified rather than guaranteed?
Do candidates differ in strategic angle and include a usable validation rule?
ChatGPT