Methodology
How Aucited measures AI search visibility
Last updated
Aucited runs an agreed set of buyer questions in real browser sessions on supported AI platforms every week, records each full response, its citations and a screenshot, and reports the findings monthly by email, with an action list and the limits stated.
What does Aucited measure?
Six signals, each recorded separately and tied to a recorded answer.
| Signal | What it means |
|---|---|
| Mention | The brand is named in the answer. It says nothing about whether the answer favors the brand. |
| Recommendation | The answer explicitly advises choosing the brand. A stronger signal than a mention. |
| Citation | The answer links or references a source URL. A citation is not an endorsement. |
| Competitor presence | Which competing brands appear in the same answers. |
| Sentiment and sources | What the answer says that is positive or negative, and the sources it comes from. |
| Trend | How each result changes across weekly checks and monthly reports. |
How does a check work?
The same prompts run every week in a real browser, and everything seen is kept as evidence.
- We agree a set of buyer questions (prompts) with you, based on your brand, market, audience and competitors.
- The prompt set is then fixed, so results can be compared over time. Any change is documented.
- Each week, we run the prompts in real browser sessions on the platforms in your plan.
- For every run we record the full response, any cited URLs and a screenshot of the session.
- We extract the findings (mentions, recommendations, citations, competitors, sentiment) from the recorded answers.
- You receive a report by email each month, with an action list. The next report shows what changed.
Why browser sessions?
Because browser sessions record what people actually see, and an API response can differ from that.
Consumer apps and search pages add features such as web search, citations and product-level instructions. Capturing the user-facing result lets you inspect the actual answer and its sources. API-based tools have their own strengths, mainly scale and frequency. Read the full comparison.
What evidence do I receive?
For each tracked answer: the full response, cited URLs where shown, and a screenshot of the session.
- The prompt, platform and week of every run
- Screenshot evidence you can download each month
- Screenshot files are automatically deleted after 30 days
- More extensive evidence collection on the Research plan
How many responses does each plan track?
Estimated responses = prompts × platforms × 4 to 4.5 weekly checks per month.
| Plan | Prompts | Platforms | Estimated responses |
|---|---|---|---|
| Starter | 20 | 2 | 160–180 / mo |
| Growth | 30 | 3 | 360–405 / mo |
| Research | 50 | 4 | 800–900 / mo |
What are the limits of these results?
Results describe what was observed in the tests we ran. They are not a guarantee or a census.
- AI answers can vary by time, session, model and context. Results describe what was observed in the the tests we ran.
- A report is not a universal measure of every possible answer, and it does not claim to reproduce every user’s experience.
- Not every platform shows every signal. For example, Google AI Overviews do not appear for every query, and some platforms show no citations. Reports note what each platform shows.
- Estimated response counts are ceilings. Real counts can be lower when a platform returns no answer for a prompt.
- No one can guarantee a brand will be mentioned in every answer. Aucited shows what to improve and tracks whether visibility is improving.
What does Aucited not do?
- Daily or real-time tracking. Checks are weekly.
- A live customer dashboard. Reports are delivered by email.
- SEO fixes or website changes. You receive an action list and make the changes.
See the evidence for your brand.
Request a sample report to see the format, the screenshots and the action list.