Know which answers changed, and why they matter
AI visibility monitoring records how a defined set of questions represents your business over time. A useful report shows the underlying answers and citations, not just a score. FoundUp combines agreed monitoring with website and content work so findings can become specific actions.
Start with questions a potential customer would ask before they know your name. Keep questions about FoundUp or your own brand in a separate group for factual accuracy checks. A branded answer is not evidence that an unfamiliar customer can discover you.
Define the observation before counting it
| Measure | What it means | What it does not prove |
|---|---|---|
| Brand mention | The answer names the business or an agreed unambiguous variant. | That it recommends the business or sends traffic. |
| Owned citation | The answer links to a page on the business's own domain. | That a reader clicked the link. |
| Answer accuracy | Checked statements agree with a dated, approved source. | That every answer on the platform is accurate. |
| Referral and enquiry | A measurable site visit or completed enquiry, subject to attribution limits. | That all AI-assisted discovery is visible in analytics. |
Calculate mention and citation rates over successful, comparable observations. Report failed runs separately. Keep question sets and branded versus non-branded results separate when comparing periods.
What a monitoring scope needs
- The exact prompt list, intent and target page for each question.
- Named platforms or APIs, market, language and available execution settings.
- Observation frequency, reporting period and the treatment of errors or missing answers.
- Saved answers, cited URLs and the rule used to identify a mention.
- A change log connecting factual corrections and published pages to later observations.
- An owner for each recommended content or technical change.
Our public observation record documents a DataForSEO API sample from this website. Consumer interfaces, logged-in sessions and other model versions can produce different results. The record is not a live benchmark.
Choose an allowance that fits your questions
The published plans list prompt and keyword allowances, priority tracking, regions, competitor comparisons and reporting. Review the current pricing table for the exact Growth and Pro limits. Confirm platform coverage and the prompts included before starting; a count alone does not describe the value of a monitoring plan.
If your team already runs an AI visibility product, compare the work that remains: verifying business facts, choosing questions, reviewing evidence and publishing improvements. Our buying guide distinguishes software access from implementation responsibility.
Monitoring questions
Are prompts actual search-volume data?
No. A test prompt is a question selected for an observation. Keyword-volume estimates can inform its topic, but do not establish how often people ask that full question in an AI product.
How quickly should I expect results?
Monitoring establishes a baseline and records subsequent changes. Recrawling, indexing and AI source selection have independent timelines. We do not promise a particular citation date or ranking.
Can I start with my own spreadsheet?
Yes. The prompt tracker and answer accuracy log show the fields needed to keep a useful record before choosing software or a service.