Free Scan
Guide

AI recommendation tracking

Being mentioned is not the same as being recommended. An answer can list your brand and still advise the buyer to pick a competitor. AI recommendation tracking looks at buying-intent prompts and records whether the answer actually recommends the brand.

SOV Tracker·Last updated: ·7 min read

Mentioned, recommended, cited

Three different outcomes can appear in one answer. The brand can be mentioned (named), recommended (advised as a choice for the question asked) and cited (one of its pages shown as a source). Each one answers a different question, so track them separately.

What Recommendation Rate measures

Recommendation Rate is measured on prompts labeled as recommendation or purchase-selection intent, such as "which tool should I use for…". It is the share of valid measured answers on those prompts that explicitly recommend the brand.

It is not part of AI Share of Voice. AI Share of Voice is the brand's share of valid mentions against a defined competitor set; Recommendation Rate is reported next to it.

Why AI recommends a competitor

When a competitor is recommended instead of you, the stored answer usually shows why:

  • The answer cites pages that describe the competitor's fit for the question more clearly than any page about you.
  • Your category, pricing or audience is stated inconsistently across your site, profiles and listings.
  • Independent reviews and comparisons include the competitor and not you.
  • The answer relies on old facts, such as a retired plan or feature, that no current page corrects.

Measure on a comparable scope

Use a fixed set of buying-intent prompts, the same market and language, and dated measurement windows. One answer is one observation; wording and choices vary between runs. A refusal or a failed request is recorded as its own state, never as a visibility zero.

Inputs that can help

Recommendation work is mostly about making the right fit easy to state:

  • Pages that say plainly who the product is for, what it costs and how it differs, with dates.
  • Consistent entity facts across your site, profiles and directory listings.
  • Accurate presence in independent comparisons that answers already cite.

How SOV Tracker tracks recommendations

SOV Tracker labels tracked prompts by intent and reports Recommendation Rate on recommendation-intent prompts, with the dated answer kept as evidence. Paid plans measure ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews weekly, plus a daily Priority Pulse on ChatGPT and Google AI Overviews. The Free scan is ChatGPT only and is a Discovery baseline.

Frequently asked questions

How do I know if ChatGPT recommends my brand?

Ask the buying-intent questions your customers ask, without naming your brand, and record whether the answer explicitly recommends you. Repeat on a fixed prompt set and dated windows, because single answers vary.

Is a mention a recommendation?

No. A mention only means the brand was named. A recommendation means the answer advised it as a choice for the question asked.

Is Recommendation Rate part of AI Share of Voice?

No. They are separate metrics reported side by side. AI Share of Voice is mention share against defined competitors; Recommendation Rate is measured only on recommendation-intent prompts.

Can the Free scan measure recommendations?

The Free scan is ChatGPT only, with five automatically generated non-branded prompts, and reports Discovery Visibility. Recommendation Rate needs tracked, intent-labeled prompts in a paid plan.

Start with a dated baseline

Run the Free scan for a ChatGPT Discovery baseline. Paid plans add intent-labeled tracked prompts, configured competitors and weekly measurement across all five engines.

Run Free Discovery Scan