Track the questions
your buyers ask.

Prompt Analysis tracks a defined set of buyer questions every day. Choose your own prompts or generate a starting set, then see how AI recommends your company over time.

Request your free report

Prompt AnalysisLaptop buying journey

Daily tracking

Active prompts40

AI platforms4

Responses per day160

Start with the questions your sales team hears every day.

3 prompts from a 40-prompt set

Apple recommendation rate75%

What are the best laptops for a 100-person company?

Last 7 daily runs+25 points

AcrossChatGPTGeminiClaudePerplexity

Choose the prompts
that matter.

Start with a fixed number of prompts that covers your buying journey. The set stays consistent, so daily changes reflect the answers to the same questions.

Choose your prompts.

Bring questions from sales calls, customer research and search intent. Write the exact wording you want to track, then group prompts by product, audience and buying stage.

Category discovery and product research
Competitor alternatives and comparisons
Purchase criteria and product requirements

Generate a starting set.

Use your company website, category and competitors to build candidate questions. Review the wording, remove duplicates and approve the final set before tracking begins.

Specific questions grounded in your offering
A balance of branded and unbranded prompts
Your approval before a question is tracked

Track your prompt set daily.

Each active prompt is run across the selected AI platforms daily. Results stay attached to the prompt, platform and run date, creating a history you can compare.

40approved prompts

4AI platforms

160responses per daily run

A 40-prompt configuration, with one response per prompt on each of four platforms.

Run

Keep the exact prompt text, selected market and language consistent. Log the model identifier and timestamp with each response.

Extract

Identify brand mentions, explicit recommendations, ranked positions where present, and source URLs returned in the answer.

Compare

Compare the latest completed run with previous days. Separate movement in your brand’s visibility from changes in competitors and citations.

Measure recommendations,
sources and accuracy.

Read the signals behind each response, then roll them up across your prompt set. Keep every score connected to the question that produced it.

Recommendation rate

How often an answer explicitly recommends your company, rather than simply mentioning it.

recommending responses ÷ valid responses × 100

Competitive position

See which brands enter the same answers and where yours appears in an ordered shortlist. An unranked mention is kept separate from a ranked recommendation.

Citation coverage

Inspect the sources cited alongside recommendations. Track recurring domains and pages, including sources that mention competitors but leave you out.

Answer accuracy

Review product claims against your reference information. Flag outdated or unsupported details so the underlying source can be investigated.

Make daily results comparable.

AI answers can vary from run to run. Read daily changes alongside a rolling trend, and inspect the underlying responses before treating a movement as a lasting improvement.

Version your set. Changing a prompt creates a new comparison baseline. Keep the previous wording and results available for context.
Keep failed runs separate. A missing or failed response is not a zero score. Compare rates using valid responses and show the coverage behind them.
Separate platform changes. Record model changes and avoid combining different model configurations into an unexplained trend.

Start with the questions
that matter to your business.

Find out where your company stands in AI answers.