View original page

See how Beacon measures AI visibility

Beacon

Separate website readiness from what selected AI answers actually say about your business.

Read the measurement guideExplore team workflows →

Illustrative interpretation

Six answers completed for one API model. None names the brand. The result is zero mentions in those six sampled answers, not six failed calls and not proof that no customer can find the business. If one answer failed, show five completed and one unavailable; do not count the unavailable answer as a negative observation.

Read the full example →

Administrator previewBeacon’s website overview and illustrative sample report are public. Private reports and workspace features require an authorized Beacon workspace with active access and the appropriate role. Free or Pro Prompeteer accounts alone don’t grant Beacon workspace access.

Website readiness and answer visibility need separate measurements

Website readiness checks whether public content can be retrieved and understood. Answer visibility examines completed answers to a defined question set, looking for brand mentions, citations and explicit recommendation positions.

This public guide explains the method and does not run model calls. The illustrative sample report is public; private reports and workspace features require an authorized Beacon workspace with active access and the appropriate role.

Measure What it can establish
Website readiness Observed technical and content conditions on the assessed pages
Sampled API answers What the saved answers said for those questions and models
Saved comparisons Differences within a compatible question and model scope

Read the denominator before the percentage

A zero in a completed sample and an unavailable observation mean different things.

Illustrative interpretation

Six answers completed for one API model. None names the brand. The result is zero mentions in those six sampled answers, not six failed calls and not proof that no customer can find the business. If one answer failed, show five completed and one unavailable; do not count the unavailable answer as a negative observation.

API observations are not consumer search rankings

A model API route is not identical to a personalized consumer ChatGPT, Gemini or Meta experience. Grounded answers and model-memory answers also have different evidence properties. Keep the model, question set, collection date and sample method visible.

No readiness score or optimization service can guarantee inclusion, a ranking position or a model’s next answer.

Common questions

Can I run a report from this page?

No. This is a public guide. Private reports and workspace features require an authorized Beacon workspace with active access and the appropriate role. Free or Pro Prompeteer accounts alone don’t grant Beacon workspace access.

Is readiness the same as being recommended?

No. Readiness measures observed website conditions. A recommendation must be evidenced in an actual saved answer.

Sources and review

Published by Prompeteer. See the people responsible for the product and contact the team with corrections. The linked workflow artifacts describe exactly what was checked; illustrative examples are not customer results.

Reviewed November 11, 2026

Continue your workflow

Read the evidence behind Beacon’s metrics →

Inspect Beacon’s AI visibility denominators, citations, sentiment quotes, recommendation positions and missing evidence. Read a worked example with reproducible calculations.

Turn a Beacon finding into a checkable action →

Use Beacon findings to decide what to change on your site. Define the evidence, owner, action and verification before claiming improvement.

Make unsupported claims easier to catch →

Use relevant sources, explicit uncertainty, verification and human review to reduce unsupported AI claims. Learn what prompts alone cannot guarantee.

Start with the approach.

Read the measurement guideExplore team workflows →


Source: https://prompeteer.ai/ai-discovery-agent