Payments made in the preview are in test mode.
MentionGlow

Our method

A measurement built for answers that constantly change.

MentionGlow tracks realistic local questions across Google AI and major AI assistants, repeats observations, and measures your presence over 28 days instead of drawing conclusions from a single answer.

Google AI Overview · AI Mode · ChatGPT · Gemini · Claude · Perplexity

Chapter 1 · What we observe

Local questions, observed in their real context.

01 · The questions

Questions built for your business.

They are generated from your services and city, then approved by you during setup. You remove any that do not apply.

Local search · Denver

Illustrative examples: Which Italian restaurant in Denver works for a group of eight?; Where can I buy second-hand clothing in Denver?; Which plumber in Denver is reliable for an emergency?

02 · The rhythm

Repeated observations, not a one-off snapshot.

MentionGlow repeats observations across supported AI-search experiences while keeping the local context. Answers can vary from one experience to another.

Google AI Overview, AI Mode, ChatGPT, Gemini, Claude, Perplexity — Denver

Repetition reveals what one answer cannot show.

Chapter 2 · Why repetition matters · 03 · Consistency

Less than 1 in 100

Why not every day?

of recommendation lists were identical from one answer to the next in a study of nearly 3,000 responses. This supports tracking repeated presence rather than a single rank.

Source: Search Engine Land, January 28, 2026 — SparkToro and Gumshoe research

04 · The measurement

One answer, one sample. A 28-day trend.

Each valid answer is one observation. Together, observations from the last 28 days show the share where your business is recommended.

Illustrative examples

28 days

Illustrative example: 12 recommendations out of 40 answers = 30%

RecommendedNot recommended

Illustrative example of 40 answers, including 12 recommendations, producing a 30 percent rate over 28 days.

Failed answers and manual checks (“Check now”) do not count.

Chapter 3 · What counts

A strict reading of recommendations and their sources.

05 · Detection

Mentioned does not mean recommended.

We recognize your business by its name, phone number and website.

If the AI places a business with the same name in another city, the name alone is not enough: the phone number or website must also match.

A business that is only mentioned (“X exists, but I recommend Y”) does not count as recommended.

“For this need, I recommend Boreal Bikes. Summit Spoke also operates in Denver, but it is not my first choice.”

Illustrative example: one business is recommended and another is only mentioned, so it does not count in the measurement.

06 · Sources

We also look at where the answers come from.

When an AI-search experience provides sources, MentionGlow records the cited domains so you can see which sites keep appearing in your market’s answers.

Illustrative examples

Illustrative example of source categories MentionGlow may identify: your website, a directory, media, a local authority or a competitor.

A cited source does not automatically mean it caused the recommendation.

Chapter 4 · Interpreting with care

The first weeks establish your starting point.

MentionGlow waits for enough observations before interpreting a change. When the measurement method changes, new data also needs time to accumulate before comparisons are meaningful.

Baseline in progress

Baseline established

08 · Limits

Measuring a change is not proving a cause.

MentionGlow shows what changed after your actions. It does not prove that an action caused the change.

Our method

See how your business appears in AI search.

Start with a free scan for your business and city.

Try an example: