What the engines say when your buyers ask, and what to do about it

Bluemoon Intelligence measures your visibility per market and engine, weights every question by buying intent, ranks the sources the engines rely on, tracks the competitors they name, and turns the gaps into briefs that carry their evidence with them.

Every answer, as the engine wrote it

One buyer question, put to every engine, every day, from inside each market.

United States · not personalised · Sep 8, 2026

What's the best waterproof hardshell jacket for alpine climbing?

For alpine climbing, the most recommended hardshells are the Alpsteel Ridgeline AR for durability, the Northlark Solano as a value pick, and the Ridgewell Cirrus Pro as a lighter option. Reviewers at GearBench rate the Ridgeline AR highest for sustained bad weather.

Ggearbench.examplequotedTtrailnotes.exampleconsultedAalpstore.exampleconsulted

Illustrative example · fictional answers. “Quoted” and “consulted” are reported by the engines that expose them; the others show only what they quoted.

The Intelligence modules

Visibility, competitors, prompts, citations, intent, alerts, agents and actions: what each one answers, drawn.

Visibility

The share of answers that name you, per engine and market, with the count next to every rate.

Illustrations, not screens. The product itself we show you in a demo.

Visibility you can trace back to answers

A score per market and per engine, built from stored answers. Whether you were named or recommended, and how citations move over time.

Illustration, not a screen: your share of answers, how it moves, who is cited beside you, and the sources behind it.

  • A score per market and engine

    Answers are collected from inside each country you sell in, so Germany and the United States are two measurements, not one. Every score shows the number of answers behind it.

  • Named is not recommended

    Each brand in an answer gets a role: recommended, listed, mentioned, caveated or absent. Visibility counts every name; recommendation rate counts only the first role.

  • Citations over time

    Which of your pages the engines cited, how often, and how that changes week to week, shown next to the previous period.

Not every question is worth the same

Every tracked question carries a weight for how close it sits to a purchase, so a gap on a buying question ranks above a gap on a definition.

Transactional
The buyer is choosing what to buy or where to buy it. A gap on these questions is the first thing to fix.
Comparison
Two or more options side by side. Whether you are recommended, listed or named with a caveat matters most here.
Informational
How something works or what to look for. Useful for being present early in a decision, and weighted lower.

The sources the engines rely on, and how to reach them

Every domain and URL cited in your answers, ranked by how often each engine used it across your questions. Then the practical part: whether you can buy a place there, pitch one, or have to earn it.

  • Run ads

    The site sells advertising. Bluemoon names who sells it and links to the public record it relied on, so your media team can buy directly or through the sales house.

    Ad-targeting sources are included on Growth.
  • Pitch

    No inventory for sale, but an editor who publishes reviews, comparisons or listings. The brief names the page the engines cite and how many of your questions it appears in.

  • Earned

    Competitor pages, retailers you already list with, communities. There is no shortcut here, and the product says so rather than suggesting one.

Competitors, as the engines see them

The brands you chose to compare against, and the ones the engines keep naming that you did not.

  • The ones you chose

    Named when you set up the workspace, up to your plan's allowance. Each is measured on every question, engine and market you track, with the same roles as your own brand.

  • The ones the engines detected

    Every brand named in a stored answer is recorded, whether or not it is on your list. When one keeps appearing, the product suggests adding it and shows the answers it was found in.

Briefs that carry their evidence

A gap becomes a brief your content, PR or listings team can act on: a page to publish, a placement to pitch, a listing to correct. Each brief names where it came from.

  • The questions

    Which tracked questions the gap was observed on, with the answer count behind each.

  • The engines

    Which engines showed the gap and which did not, so a brief for one engine is not mistaken for a brief for all of them.

  • The sources

    The cited pages the brief points at, and how many of your questions they appear in.

  • The intent weight

    Why the brief ranks where it does, and whether that weight was anchored on ad prices or read from the wording.

Briefs with measured impact, which track the same questions after the work is done, are included from Core.

One answer, worked through

An illustrative example: one question, one engine, one market, and what Bluemoon does with the answer.

ChatGPTUnited States · Sep 8, 2026 · not personalisedObserved answer

What's the best waterproof hardshell jacket for alpine climbing?

For alpine climbing, the most recommended hardshells are the Alpsteel Ridgeline AR for durability, the Northlark Solano as a value pick, and the Ridgewell Cirrus Pro as a lighter option. Reviewers at GearBench and TrailNotes rate the Ridgeline AR highest for sustained bad weather, while the Cirrus Pro is praised for mobility on technical routes.

Sources the answer relied on
  • Ggearbench.example3 of 6 engines
  • Ttrailnotes.example2 of 6 engines
  • Aalpstore.example2 of 6 engines

Brands in this answer

  • AAlpsteel Recommended
  • NNorthlark Listed
  • RRidgewell YouMentioned
  • VVantera Not named

The gap, across 6 engines

  • Alpsteel recommended5 of 6
  • Ridgewell named3 of 6
  • Ridgewell recommended0 of 6

Proposed action

Get the Cirrus Pro into GearBench’s alpine hardshell review.

That page is cited by 3 of 6 engines for this question and for 14 other questions you track.

High priorityTransactional intentOwner: PR
Illustrative example · fictional data
  1. The question is asked and the answer is stored

    “What's the best waterproof hardshell jacket for alpine climbing?” goes to ChatGPT from inside the United States, logged out. The full answer is stored with the time it was collected and the three sources it relied on.

  2. Brands and their roles are read from the answer

    Alpsteel is recommended, Northlark is listed as a value pick, Ridgewell is mentioned as a lighter option and Vantera is not named. The roles come from the answer text, not from a guess about it.

  3. The gap is counted across engines

    The same question on every engine: Alpsteel recommended in 5 of 6, Ridgewell named in 3 of 6 and recommended in 0 of 6. Counts rather than rates, because six answers are too few for a percentage to mean much.

  4. A brief is proposed with its provenance

    Get the Cirrus Pro into GearBench's alpine hardshell review. That page is cited by 3 of 6 engines for this question and for 14 other questions in the workspace. Transactional intent, high priority, owner: PR.

Illustrative example · fictional data. Not a customer result.

Observed or inferred, never mixed

Every finding keeps its question, engine, market, collection time, the stored answer and its sources. Counts from stored answers are marked observed. Anything derived, like an intent weight, is marked as an estimate with its inputs listed.

A dash means it was not measured. Zero means the engines were asked and never named you.

Finding

Ridgewell is named in 3 of 6 engines for alpine hardshell questions and recommended in 0.

Observed over 12 questions · 144 answers · Aug 10–Sep 7, 2026
Question
best hardshell jacket for alpine climbing
Engine
ChatGPT, web, not personalised
Market
United States, collected from inside the market
Collected
Sep 8, 2026, 06:12 UTC
Answer
#a7f3-2026-09-08-us-chatgpt · stored in full
Sources in answer
gearbench.example · trailnotes.example · alpstore.example
ObservedCounted from stored answers.EstimateNever shown as a rate. Where something was not measured you see a dash.

Questions about the data

What is observed and what is inferred?
Anything counted from stored answers is observed: whether you were named or recommended, which sources were cited, which competitors appeared, how often. The intent weight on a question is an estimate, and sentiment by aspect is read from the answer text by a model; both are labelled as such wherever they appear. Observed counts and estimates are never combined into one figure.
How often does the data update?
Every tracked question is collected once a day on every engine and market in your plan, and every answer is stored in full. You can also run a scan on demand within the daily allowance. A failed collection is recorded as a failure, not filled in with a guess.
Which engines are covered?
ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Google AI Mode and Microsoft Copilot. Claude is collected through Anthropic's API with your own key on every plan. Starter collects three, Core five, and Growth all seven.
How are competitors detected?
Every brand named in a stored answer is recorded, whether or not it is on your list. When a brand you did not list keeps appearing for your questions, the product suggests it and shows the answers it was found in. You decide whether to add it, up to your plan's allowance.
How should I read the intent weight?
As an ordering, not a measurement. The weight is an estimate, labelled as one, and kept apart from observed counts. Use it to decide which gap to work on first, and rely on the observed counts for what actually changed.

Start with your questions

Seven-day free trial. Pick a market, keep or edit the generated questions, and read the first stored answers within the hour.