Week-over-week changes
See the prior and current values for visibility score, answer mention rate, and share of voice, including a clear first-week state.
AI Visibility Reporting
Turn a week of AI answers into the next clear action
Compare core visibility metrics with the prior week, expose topics where competitors appeared and you did not, and keep every recommendation grounded in recorded answers and citations.
Reports summarize your configured prompts and supported API providers. They do not estimate all ChatGPT or Claude usage, traffic, leads, or revenue.
Product scope
What this page covers
- Comparison
- Week over week
- Visibility score, mention rate, and share of voice
- Diagnosis
- Wins + topic gaps
- Examples link back to the underlying answer record
- Planning
- Sources + actions
- Recommendations are constrained to computed facts
Reporting without dashboard drift
A metric is only useful when it changes the work queue
AI visibility can generate a large volume of answer data: multiple prompts, two providers, daily or weekly runs, mentions, ranks, sentiment, and sources. Looking at every transcript is valuable for investigation but inefficient as a weekly operating rhythm.
The weekly report compresses that record into three decisions: what moved, where a competitor won, and what topic or source deserves attention next. Core changes and gaps are calculated in code. AI writes concise explanatory prose from those supplied facts and is explicitly instructed not to invent numbers or domains.
The report is a prioritization layer over stored evidence: each metric comes from recorded answers, and competitor examples lead back to the transcript.
How it works
From scheduled observations to an evidence-backed weekly brief
Reports become more informative as a stable prompt set accumulates successful runs across adjacent weeks.
- 1
Accumulate the week's answer set
Scheduled or manual monitoring stores successful answers with their prompts, topics, providers, mentions, rank, sentiment, and returned citations.
Output: A bounded weekly evidence set
- 2
Compare three core metrics
Calculate visibility score, the percentage of answers mentioning your brand, and share of voice for the current and previous week, then label each change up, down, or flat.
Output: A consistent week-over-week scorecard
- 3
Find competitor wins and topic gaps
Identify example answers where a configured competitor appeared and your brand did not. Group broader gaps by topic and retain the prompts, competitors, and external cited domains involved.
Output: Specific losses instead of a vague decline
- 4
Translate facts into a publishing priority
Generate concise prose constrained to the computed changes, wins, gaps, and domains. Open the linked answers and sources before committing resources.
Output: A short, reviewable action list
Inside the report
The information needed for a weekly search-and-content review
The report keeps measurement, diagnosis, and action in one place while preserving links back to evidence.
Competitor-win examples
Name the competitor, topic, provider, and one supporting answer where it was recommended and your brand was absent.
Topic gap analysis
Group prompts where competitors appeared and your brand did not, then retain the competitor names and cited domains involved.
Publishing targets
Surface external domains returned in gap answers as places to investigate for coverage, contribution, or content-pattern research.
Grounded report prose
Use AI for concise headlines and recommendations only after the metrics and factual gap inputs have been computed in code.
Report history
Keep weekly reports by ISO week and browse recent periods instead of replacing the previous summary with the latest one.
Report structure
A brief designed to answer four questions
Each section narrows a large answer set into a decision your team can review.
A report is not generated when there is no current-week stats record. The first report can establish a baseline but cannot provide a meaningful prior-week comparison.
Product record
- What moved?
Three metric deltas
Visibility, mention rate, and tracked-brand share of voice
- Who won?
Competitor + topic + engine
With an example answer ID for verification
- Where is the gap?
Topic + prompts + cited domains
Only where competitors appeared and your brand did not
- What next?
Grounded publishing action
Written from the computed facts, then reviewed by a human
Operating rhythm
Give each weekly review a narrower, better agenda
01
Content planning
Choose the next comparison, use-case page, data asset, or refresh from a topic where competitors repeatedly appear and your brand does not.
02
Editorial outreach
Review external domains cited in competitor-gap answers and determine where a legitimate expert contribution or earned mention is possible.
03
Executive visibility updates
Share a compact trend and risk summary while keeping the scope clear: a monitored prompt set, not universal AI market share.
A weekly report is a decision aid, not an attribution model
The report narrows attention to observable patterns. It cannot prove that one page, campaign, or source caused a generated answer to change.
- A first week has no prior-period evidence, so changes are shown as a baseline rather than a meaningful trend.
- The report covers only successful answers from the configured prompts, topics, competitors, and OpenAI or Anthropic providers in the project.
- Publishing targets are domains cited in relevant gap answers. Inclusion does not guarantee editorial access, a link, a future citation, or improved visibility.
- Recommendations organize the next investigation; they should be reviewed against the raw answers, source quality, business priorities, and available expertise before execution.
Common questions
Questions about ai visibility reporting
Clear answers about scope, evidence, and what the current product actually does.
What appears in a weekly AI visibility report?
The report includes changes in visibility score, answer mention rate, and share of voice; example competitor wins; topic gaps where competitors appeared and your brand did not; external domains cited in those gaps; and concise recommendations grounded in those computed facts.
When are reports generated?
The product is set up to generate reports automatically each Monday for the prior monitoring week. The reports view also includes a control to generate the current week's report when enough data has been recorded.
Are report metrics generated by AI?
No. Metric deltas, topic gaps, and competitor-win inputs are calculated in code from stored answer records. An AI model is used only to write concise headline and recommendation prose from those supplied facts, with instructions not to invent numbers or domains.
What happens in the first week?
The current values establish a baseline, while change notes indicate that there is no prior week of data. Week-over-week movement becomes more useful after a stable configuration has produced successful answers in adjacent weeks.
Can I verify a competitor win in the report?
Yes. A competitor-win item retains an example answer ID and the report interface links it to the answer record, where you can review the prompt, provider, response, mentions, sentiment evidence, and returned citations.
Does the report measure leads or revenue from AI assistants?
No. It reports visibility inside your recorded OpenAI and Anthropic answer set. It does not track every consumer session or attribute website visits, leads, pipeline, or revenue to an AI recommendation.
Continue with the next useful page
AI prompt monitoring
Control the buyer-question set and cadence that supply the evidence summarized by weekly reports.
Open pageAI visibility report template
Use a manual, vendor-neutral reporting structure before or alongside automated weekly reports.
Open pageAI citation source tracking
Inspect the domain-level source data that feeds publishing targets inside topic-gap reports.
Open pageReplace scattered answer checks with a weekly decision rhythm
Monitor a consistent prompt set, preserve the evidence, and let each report point your team toward the gap worth reviewing next.
Start weekly AI reporting