Prompts / Prompting & AI Workflows

ChatGPT prompt for an unbiased review

This prompt creates a critical review that distinguishes supplied evidence from uncertainty and marketing claims. Use it to compare a product, service, provider, or other option without asking AI to manufacture certainty.

PromptOpen ChatGPTOpen Claude
Write a balanced, evidence-labeled review of [ITEM OR SERVICE] for [READER AND USE CASE]. Base it only on [SOURCES OR NOTES], plus these evaluation priorities: [EVALUATION CRITERIA]. My constraints are [BUDGET OR CONSTRAINTS].

Do not invent firsthand experience, test results, customer feedback, prices, specifications, credentials, quotes, or current availability. Do not treat marketing copy, a star rating, or a single anecdote as proof. If a required input is missing, ask up to 3 clarifying questions.

Produce 450–650 words with these sections:

- Bottom line: a conditional recommendation in 2–3 sentences stating who it suits, who should avoid it, and why.
- What is supported by the evidence: 4–6 factual claims, each labeled with its source type.
- Strengths and limitations: two to four of each, tied to the stated criteria rather than vague praise or criticism.
- Tradeoffs and alternatives: explain what a buyer gives up by choosing it; compare [COMPARISON OPTIONS] only on evidence provided.
- Unknowns and verification steps: list information that could change the conclusion and exactly what to verify before buying, booking, or relying on it.

Use calibrated language: “the supplied documentation states,” “one review reports,” or “not established by the available evidence.” Separate fact, inference, and opinion explicitly. Before answering, check that favorable and unfavorable evidence receive comparable scrutiny and that every factual claim is traceable to a supplied source.

Fill in

PlaceholderWhat to enterExample
[ITEM OR SERVICE]Enter the product, service, professional, or option being reviewed.Northline Carry-On suitcase
[READER AND USE CASE]Describe who needs the review and the job they need the item to do.A frequent business traveler taking two- to three-day domestic trips
[SOURCES OR NOTES]Paste product details, reviews, test notes, policies, or source excerpts.Manufacturer: 39L, 7.1 lb, polycarbonate shell, five-year warranty. Two owner reviews mention smooth wheels; one says the handle loosened after eight months. Return policy: 30 days, unused only.
[EVALUATION CRITERIA]List the criteria that matter most, such as durability, service quality, safety, or total cost.Wheel durability, overhead-bin fit, warranty clarity, and total cost
[BUDGET OR CONSTRAINTS]Enter practical limits such as budget, location, compatibility, allergies, or timing.Under $250; must fit typical US domestic overhead bins
[COMPARISON OPTIONS]List alternatives to compare, or write none.Travelpro Maxlite 5 and Away The Carry-On

How to use

  1. Paste the prompt with source excerpts rather than asking the model to rely on general reputation.
  2. Set criteria before reading the conclusion so the review reflects your actual decision.
  3. Check the “unknowns” section and verify any fact that could change your purchase or safety decision.
  4. Follow up with: “Turn this into a weighted comparison table using only the evidence above; leave unsupported cells blank.”

Variations

Beauty product

Use when assessing a cosmetic or skincare product for a particular routine.

Variation
Assess [PRODUCT] for [SKIN TYPE OR HAIR TYPE] and [GOAL] using only [INGREDIENT LIST AND SOURCES]. Identify ingredients with known functions, likely irritation or compatibility concerns, fragrance or allergen considerations, and claims not established by the supplied evidence. Do not diagnose conditions or promise results. Produce a 350-word review with suitable users, users who should patch-test or consult a dermatologist, evidence limits, and a comparison with [ALTERNATIVE]. Check that product claims are not presented as clinical proof.

Car comparison

Use when comparing vehicles against real ownership needs.

Variation
Compare [VEHICLE A] and [VEHICLE B] for [DRIVER PROFILE] using [SPECS, PRICES, AND SOURCES]. Evaluate purchase price, ownership-cost assumptions, cargo and passenger needs, safety features, warranty, range or fuel economy, and known limitations. Separate manufacturer claims from independent evidence. Return a decision table, a conditional recommendation, and questions to verify on a test drive or dealer quote. Do not invent current incentives, reliability data, or safety ratings. Check that comparisons use equivalent trims and model years where stated.

Provider shortlist

Use when reviewing a professional service provider before a consultation.

Variation
Create a neutral shortlist review of [PROVIDER OPTIONS] for [NEED] in [LOCATION] from [SOURCES]. Compare stated qualifications, scope of service, access, fees if supplied, policies, and evidence limits. Do not rank providers from star ratings alone or make medical, legal, or financial judgments beyond the source material. Produce a table plus five consultation questions that help confirm fit. Check that each provider is judged against the same criteria and flag missing credentials or pricing information.

Tips

  • Set evaluation criteria before gathering evidence; otherwise a review drifts toward whichever claims are easiest to find.
  • Give anecdotes their proper weight. They can reveal failure modes worth checking, but they do not establish typical performance.
  • Compare equivalent versions, plans, trims, or service scopes; price and feature comparisons collapse when the underlying options differ.
  • For high-stakes services, use the review to prepare questions for a licensed professional, not to choose solely on AI’s conclusion.

FAQ

Can AI write an unbiased review from online information?

It can organize sources and expose uncertainty, but “unbiased” depends on the quality and completeness of the evidence supplied. Ask it to label source types and unknowns.

Should I include negative reviews?

Yes, especially recurring complaints with details. Include positive and negative evidence so the model can assess tradeoffs instead of producing praise.

Can this prompt review doctors or financial advisers?

It can compare publicly stated qualifications and policies, but it cannot verify professional quality or replace a licensed medical, legal, or financial assessment.

Related prompts

All Prompting & AI Workflows prompts →

Get a few prompts like this every week

Something Big is a free AI newsletter read by 50,000+ professionals. One email a week with prompts and tools that work, plus what changed in AI and what to do about it.