AI review summary: what assistants tell your buyer about your reviews
Reviewscope shows the AI review summary your buyer hears: what ChatGPT, Claude and Gemini say about your brand's reviews, platform by platform, with every rating marked as an unchecked claim and the invented ones flagged.
Three assistants, three ways of asking. Their claims, kept apart from platform facts.
Before a buyer opens G2 they ask an assistant. Reviewscope puts three questions to each of three assistants — one naming no platform, one naming all twelve, one naming three platforms we invented — and shows what came back, including the ratings they made up.
A free allowance per visitor, no account needed.
This tool reports the AI review summary assistants give. It does not report your real rating.
Every rating on this page is a claim from an AI assistant and may be inaccurate, incomplete or invented.
Questions
What is an AI review summary?
Why not just read my G2 or Capterra rating directly?
How do you know a platform an assistant named does not exist?
Is a rating an assistant remembers my real rating?
Is a run free of charge?
How to use Reviewscope
- 1 Name a brand or websiteA company name or its domain. A domain also lets us read the rating the brand states itself.
- 2 Compare each AI review summarySee what each assistant says about your reviews, kept separate from platform facts.
- 3 Use the tabs and the filterSwitch between unprompted, when asked and invented. Toggle numerical claims only.
Why ask an assistant about reviews at all
A buyer comparing tools used to open two review sites and read. Now a growing share of them asks an assistant first, and what they hear is an AI review summary rather than the reviews: a platform name, sometimes a score, delivered in the same confident register as everything else. Whether that summary resembles your reputation is a question about the assistant, not about the platform.
That question splits three ways, which is why this tool asks each assistant three times. Without a platform named, an assistant reveals which sources it reaches for on its own — the ones that will reach your buyer unprompted. With every platform named, it reveals how willing it is to produce a number it does not have. With three platforms we invented, it shows whether it will rate a place that is not there — the clearest finding here, and the one we can prove rather than report.
None of the three is a rating source, and this page never pretends otherwise. What a platform actually scores you is on that platform, and it belongs to them. What an assistant says you score is a different fact about a different system — and it is the one your buyer meets first.
Who it is for
- Marketing and brand teams who need to know which AI review summary a buyer hears before reaching a review site — and whether any of it is invented.
- Whoever runs a review programme on G2, Capterra or Trustpilot and has to decide which platforms the assistants actually repeat.
- Founders and support leads handed a screenshot of an assistant quoting a rating nobody recognises, who need to know whether it is a habit of the assistant or a fact about them.
To see whether the assistants name your brand at all, start one step earlier with Brandscope. Every free tool, in the order they are meant to be used, is on the free tools page.
How to read the three questions
Each tab reads the AI review summary from a different angle, and none of them answers what your rating is.
- Unprompted
- Which platforms an assistant names with no prompting. A platform cited by all three is part of your reputation whether or not you maintain a profile there; a platform cited by none is one your buyer will not hear about from an assistant.
- When asked
- What each assistant claims once every platform is named to it. Read the count under the column before the cells: an assistant that produced eleven numbers out of twelve has told you more about itself than about you.
- Invented
- A score attached to a platform that does not exist: one of the three we invented and named in the question, or one an assistant produced on its own. One of these makes every other number in that assistant's column a claim to discount — and unlike everything else on the page, it is a thing we established rather than repeated.
What this tool will not do
- It never states a rating for your brand, and it writes no AI review summary of its own. Not an average, not a best guess, not a figure gathered from the assistants and presented as yours.
- It reads no platform's own numbers. G2's, Capterra's, Trustpilot's and the rest are theirs, and taking them would break their terms — so every number here is an assistant's sentence, shown as one.
- It calls one thing false: a rating for a platform that does not exist. Everything else an assistant said is shown unchecked, and labelled unchecked.
- It reports what it covered. Twelve platforms, three invented ones, three assistants, nine questions — and when a list is cut, the count of what was left out is on the screen.
The same questions, out loud
SpokenHow people actually ask this, and what the honest answer is.
Because a named platform invites a plausible number, and a plausible number is what you get. Run the probe: if the same assistant also produces a score for a platform that does not exist, you are looking at a habit rather than a fact about you, and that is the thing to say when somebody forwards you the screenshot.
A customer sent a screenshot of an assistant quoting a rating nobody recognises.Start from the unprompted table rather than from the platform's sales deck. A site all three assistants reach for on their own is already reaching your buyer; a site none of them mentions may still be worth a profile, but it will not be what an assistant repeats.
Deciding where a limited reviews budget goes this quarter.The two are not connected as tightly as they feel. An assistant's AI review summary comes from what it absorbed, not from today's average, and it may not name your best platform at all. The unprompted table shows which sources it is drawing on, which is where a good rating has to be before it can be repeated.
A strong score on one platform and a flat answer from the assistant.Words on this page
- Hallucination
- A confident statement a model produces that has no source behind it. A rating for a review platform that does not exist is the cleanest example there is: the shape of the answer is right and the referent is missing.
- Brand mention
- An assistant naming a company in an answer. Counted here per platform rather than per answer, because being named beside G2 and being named beside a site nobody has heard of are different facts.
- AI visibility
- How much of what an assistant says about a market includes you. Reputation is one slice of it — the AI review summary, which a buyer reaches last before deciding.
On this topic
What do models believe about our brand?
- Check
- Brandscope
Every day Competitor comparison, answer by answer. How the product works
Other free AI search tools
- Vectorscope Paste a public URL to see its readable passages, semantic groups and the content that falls outside the retrieval cap.
- Voicescope Start with one keyword and explore the questions a person may ask an AI assistant across real-life contexts.
- Brandscope Ask three AI assistants the buyer questions around a brand, then compare their answers with the brand's own page.
- PatternScope Paste a page to see the stock phrases, repeated sentence shapes and unchecked claims a reader reacts to. Not an AI detector.
- AI answer volatility A daily reading of how far AI assistants' answers moved since yesterday, so a drop in your own visibility can be read against the day it happened on.
- All free tools