A coverage report is easier to trust when another person can open the original story and understand exactly what was checked. A headline count alone cannot show whether a mention is new, whether the same article was entered twice, or whether an AI answer actually cited it.
Shane’s AI PR gives you a local Coverage log for that recordkeeping. This guide explains how its fields, citation labels, search and exports work, then turns them into a practical reporting routine. For pitch preparation and contact research, see the media outreach workflow guide.
1. Decide what one coverage entry represents
For a straightforward report, use one entry for each published article you want to track. Open the source first, confirm that the relevant mention is there, and record the article title, outlet, URL and publication date. The tool requires an Article / mention title, but the other fields are optional. Your reporting standard should be more complete than the minimum the form accepts.
Keep the publication date separate from the date you checked it. Published date describes the story. Put the verification date in Notes / query used to verify, along with a short description of the mention and where to find your supporting evidence.
The current log does not automatically detect duplicate URLs. Two entries for the same article count as two coverage records. Before adding a story, search for a distinctive title phrase or the outlet name and inspect any matching entries. Search covers titles, outlets and notes; it does not search the URL field.
Also record the kind of coverage in Notes, such as an editorial story, contributed article, paid placement or press-release republication. If several sites carry the same release, state how you count those placements in your report. A row count cannot decide whether they represent separate editorial attention.
2. Use the three citation labels consistently
The LLM citation checked? field has three choices. Give each one a clear meaning in your workflow:
- Not checked: you have not reviewed an AI answer for this record. This is the default selection. It does not establish whether the article appears in AI answers.
- Yes, cited: you inspected a particular answer and verified that it cited the source you are recording. Save enough context to identify that answer and the linked article.
- Checked, not cited: you inspected a particular answer and did not find the citation in that check. Keep the same context so the result can be interpreted within its limits.
A brand mention in an answer and a citation to the article are different observations. If the answer names your brand but links to another source, describe that in Notes rather than treating it as evidence that this coverage URL was cited.
Include the answer engine, the exact prompt, the date of the check and a reference to a saved answer or screenshot you can review later. The tool stores the label you select; it does not retrieve AI answers, verify citations or monitor for changes. A single check cannot establish what every user or later query will receive.
3. Write a note that explains the evidence
Keep the in-tool note compact and maintain longer evidence in your usual working document. The following fictional example illustrates a record structure, not a real placement or measured result:
Campaign: Autumn workshop
Coverage type: Editorial story
Mention: Workshop format described in paragraph 4
Source checked: 2026-11-04
Citation check: Example engine; 2026-11-04
Prompt: [Exact wording used]
Result: Checked, not cited in this answer
Evidence: Internal campaign record, item 03Include a consistent campaign name in Notes if you want to find related entries later. There is no separate campaign filter, but coverage search checks the notes text. Use the same spelling each time so a search for “Autumn workshop” retrieves the intended records.
Sentiment is also a manual selection. Choose Neutral, Positive or Critical based on the actual coverage and add a short explanation when the classification could be disputed. The tool does not analyze the article’s sentiment for you.
For repeated citation checks, keep a dated history outside the tool and reference it from the coverage record. The current coverage list has a Remove control but no edit or check-history control. Adding another row for every repeat check would increase the coverage count, even when the published article has not changed.
4. Read the headline numbers as workspace totals
Coverage logged counts all saved coverage entries. LLM cited mentions counts the entries whose citation choice is Yes, cited. Both numbers describe records in this browser’s workspace, including duplicates or incomplete entries you may have added.
Typing into Search coverage narrows the visible list. It does not recalculate those headline numbers for the search results. If you search for one campaign, do not copy the totals at the top into that campaign’s report without checking the matching entries separately.
For example, imagine four saved records: two marked Yes, cited, one marked Checked, not cited and one marked Not checked. The headline numbers show four coverage records and two cited entries. The fourth record has no citation-check result. Report the breakdown explicitly rather than presenting the totals as an overall citation success rate.
A useful summary could read: “Four coverage entries recorded. Three have a documented citation check: two cited and one not cited. One remains unchecked.” Add the time window, engines, prompts and counting rules before drawing wider conclusions.
5. Export the file that contains the coverage
Use Export workspace at the top of the tool to download the JSON record. It contains the saved pitches, contacts and coverage. The Export CSV control in Pitch builder exports pitch follow-ups only, with topic, outlet, status, creation time and draft text. It does not include the coverage log.
Before relying on an export, open the downloaded JSON in a text editor or JSON viewer and confirm that the coverage entries and notes are present. Give the file a recognizable date and campaign or workspace label. Keep it somewhere appropriate for the contact information and draft material it also contains.
Data stays in this browser’s local storage. Another device or browser does not automatically receive it, and clearing stored site data can remove the records. The current interface provides an export but no workspace import or restore control. Keep the downloaded file as a readable external record without assuming it can be reloaded through the tool.
A short check before sharing the report
- Open each source URL and confirm the recorded mention.
- Check for duplicate stories and explain how republications are counted.
- Separate unchecked entries from checked results and keep citation evidence with the report.
- Use totals for the intended campaign and date range, rather than unfiltered workspace counts.
- Verify the JSON export and share only the records the recipient needs.
With those habits, the Coverage log becomes a useful index to your evidence. Every number has a clear scope, every citation label points to a specific check, and a teammate can follow the source trail without guessing.