Where AI helps.
Where you decide.

Using only the examples I provide, suggest three buyer/problem pairs. Separate evidence from assumptions. For each, give a question about a recent real experience that I can ask a potential buyer. Do not invent research. Examples: [paste anonymized notes].

Draft three small paid-pilot scopes for this problem: [problem]. My relevant skills: [skills]. For each, list the deliverable, exclusions, customer inputs, and an acceptance check I can demonstrate. Do not promise business results.

Review this pilot estimate: [tasks, hours, costs, price]. Identify missing work, revisions, sales time and support. Keep all uncertain values marked unknown. Do not set a price from invented market rates.

Write a short invitation to discuss this pilot: [scope]. Use only this proof: [evidence]. Ask about their current process before pitching. Do not invent familiarity, urgency, results or testimonials. Return one draft and two genuine questions.

Summarize these actual pilot results: [notes]. Separate observations, interpretations and unknowns. Suggest one low-cost next test for each interpretation. Do not treat silence or a tiny sample as proof of demand.

Outline a one-page site with opening, offer, process, evidence, questions and next action. Use my inputs: [brief]. Flag missing facts instead of filling them in. Keep one primary action.

Draft page copy from this approved brief: [brief]. Use plain language and concrete deliverables. Do not invent testimonials, statistics, credentials or availability. List every detail I must confirm before publishing.

Implement this approved page within my existing project: [brief]. Inspect the project instructions first. Preserve existing routes and forms. Use semantic HTML, visible focus and mobile layouts. List needed integrations; do not fake form success or deploy yet.

Create an acceptance checklist for this page: [brief]. Include navigation, keyboard focus, small screens, form validation, successful storage, error messages and confirmation wording. Distinguish checks performed from checks still pending.

Prepare a release checklist for [site]. Include live links, forms, canonical URLs, sitemap, search ownership verification and consent-based measurement. Treat account setup and indexing as unverified until there is evidence. Define one useful action to measure.

Group these real questions by intent: [questions]. Choose three possible article topics and explain what a reader would be able to do afterward. Do not invent search volume, popularity or quotes.

Create a research checklist for [question]. Separate claims requiring a source, experience I can demonstrate, and assumptions. If sources are supplied, map each claim to the exact source; do not fabricate citations.

Draft a guide using only this evidence: [notes and sources]. Answer [question]. Include a worked example, steps, common mistakes and limitations. Mark unsupported statements [VERIFY]. Do not imply I tested something I did not.

Critique this draft against its sources: [draft and ledger]. List unsupported claims, missing steps, unclear language and repeated filler. Suggest edits without inventing new evidence. Check that the call to action follows naturally from the reader task.

Adapt this approved article into one short social post, a 60-second demonstration outline and an email excerpt: [article]. Preserve caveats and source context. Give each one useful idea and one next step. Do not add claims or predict performance.

Create a comparison rubric for [buyer and task]. Include fit, limitations, total cost, effort and a no-purchase alternative. Do not choose a winner or recommend a product before evidence is provided.

Organize these notes into a comparison table: [notes and official sources]. Label each point firsthand, vendor claim or unknown. Never invent a hands-on test, rating, price or result. Flag inconsistencies for me to check.

Write a comparison using only this rubric and evidence: [material]. Explain who each option fits, who should skip it and what remains uncertain. Avoid universal best claims and unsupported earnings or performance promises.

Review this disclosure and draft for clarity: [text]. My actual relationship is [relationship]. Flag any place a reader could miss that I may receive compensation. Do not certify legal compliance; identify questions to check against FTC guidance and the program terms.

Analyze these aggregate affiliate results: [counts, costs, hours, period]. Distinguish clicks, pending commissions, approved commissions and cash received. Show arithmetic, missing data and small-sample limitations. Suggest one test without inventing a benchmark.

For this offer [offer], suggest observable signs that a business might face the problem. Give research questions, not a scraped contact list. Do not infer private characteristics or invent contact information.

Draft five discovery questions for [problem]. Ask about the last time it happened, their workaround, impact and prior attempts. Avoid asking whether they like my idea or hypothetical willingness to pay.

Draft a brief invitation based only on [verified context]. Ask one low-pressure question about [problem]. Include no invented familiarity, urgency or results. Make declining easy. Do not send or schedule anything.

Using these anonymized notes [notes], draft a follow-up with the agreed resource and next step. Do not add commitments or a sales pitch they did not ask for. If no follow-up was invited, suggest a respectful stop instead.

Group these observations [notes] into evidence, objections and open questions. Suggest one offer change and one test to evaluate it. Do not infer a conversion rate from incomplete counts or treat a compliment as a sale.

Structure these aggregate records [records] into hours, costs, activities and outcomes for [period]. Label missing values unknown and estimated values estimated. Do not treat missing data as zero.

From this log [log], separate activity, attention, intent, commitment and paid outcome. Explain denominators for any rate. Do not count my own test visits or test submissions as customers.

Audit these totals [numbers]. Show the calculation, list omitted cost categories and distinguish collected revenue from pending amounts. Include selling, delivery, revision and support time. Do not give tax or investment advice.

Compare this result [result] with my original criterion [criterion] and resource limit [limit]. Offer continue, change and pause options with evidence for and against each. Identify what would change my mind; do not guarantee an outcome.

Draft a build-in-public update using only [approved facts]. Include what I tried, effort and costs I choose to share, what happened, what I learned and the next test. Do not turn projections into results or invent a success story.