Measurement Scope & Boundaries

1. What this statement is for

A common question is: "Can you guarantee we rank #1?" Our answer is no. That is not evasion but a fact of GEO (generative engine optimisation) - AI answers are generated on the fly, unlike a fixed search ranking. This statement sets out clearly what we guarantee, what we do not, and how results are measured, so you can use it during procurement and acceptance.

2. What we guarantee (deliverable and verifiable)

(1) Real sampling: every monitoring request is a real call to the target engine, with the full answer retained and traceable - not simulated, not cached, not pre-written; (2) Transparent process: reports show the engines, questions and sampling times, and you can check each answer; (3) Consistent metrics: all indicators use one metric definition (see section 4), and reports, dashboards and the console share the same source; (4) Service fulfilment: delivery at the frequency, scope and deliverables agreed in the contract (content, distribution, reports).

3. What we do not guarantee (and why)

(1) We do not guarantee a specific rank or "being #1" - AI regenerates answers each time, so the same question can yield different answers at different moments; (2) We do not guarantee a fixed visibility rate - it is probabilistic and varies with wording, engine version and timing; (3) We do not guarantee lead or sales volume - between "being mentioned by AI" and "a customer buying" lie landing pages, sales and pricing, which are outside this service; (4) We do not guarantee competitors will not act - this is a dynamic competitive environment and their investment can change results too.

4. How results are measured: the four-layer framework

The report centre uses a four-layer framework - Exposure -> Cognition -> Trust -> Conversion - that builds up progressively; looking only at the first two layers is self-deception: Exposure (AI visibility) = share of AI answers mentioning your brand, i.e. "does AI know you"; Cognition (citation rate) = share of answers citing you as a source, i.e. "being mentioned is not the same as being trusted"; Trust (top-recommendation rate) = share of questions where AI ranks you first, i.e. "will AI recommend you"; Conversion (qualified leads) = recent qualified leads (including those from AI), i.e. "will customers come to you". Each layer comes with a health verdict and a suggested action.

5. Read the trend, not a single point

A single sample fluctuates - that is the nature of probabilistic sampling. We recommend judging by trend rather than a single point: is visibility rising over consecutive weeks, are more questions citing you, are negative or incorrect descriptions decreasing. Reports support comparison against your earliest report as a baseline - an effect compared without a baseline is meaningless.

6. A verifiable evidence chain

(1) Traceable samples: each answer keeps the engine, question, timestamp and full text; (2) Checkable sources: reports list the domains AI cited, which you can open and verify; (3) Testable website: the website GEO audit gives reproducible scoring rules and details across structure, content and citability; (4) Trackable distribution: records and receipts for each channel are available (platform-published; receipts are per platform).

7. Boundaries and disclaimers

(1) Engine behaviour changes: model versions, retrieval strategies and answer styles are controlled by the vendors, and their changes may affect monitoring results; we are not responsible for third-party engine behaviour; (2) Sampling cost: monitoring consumes AI quota, calculated per your plan, with top-ups when exceeded; (3) Nature of the data: this service outputs monitoring and optimisation guidance; final business decisions and outcomes rest with you; (4) Compliance responsibility: final compliance of content (industry licences, advertising law, platform rules) is confirmed by you; the platform provides word lists and pre-checks as an aid, not a replacement for your review.

8. Why we make no performance promise

In the AI era, any claim of "guaranteed ranking / guaranteed #1" does not match the technical facts - AI is a probabilistic model and there is no fixed position that can be locked. Instead of empty promises we offer real sampling, transparent metrics, trend-based acceptance and verifiable evidence. It will not give you an inflated number, but it lets you see where your brand truly stands in AI and keep improving from there.