Every mention found the day it publishes, quote included

What is the exact prompt?

Daily, early, before the team's morning scroll finds things first:

• "Search for new mentions of [brand], [product names] and [founder names] since yesterday: news search, the newsletters on [watchlist], podcast directories (search show notes and episode titles), YouTube titles and descriptions, and new 'best [category] tools' or '[competitor] alternatives' roundups. Exclude our own domains and social accounts."
• "For each mention, read the surrounding context, not just the sentence, and log a row: date, outlet, author, link, mention type (feature / quote / listicle inclusion / passing reference / comparison), what they actually said in one line, sentiment (positive / neutral / negative / mixed), factual accuracy (correct / minor error / significant error; quote the error), and a reach proxy (stated subscriber count, or 'unknown')."
• "Flag to [#comms] immediately, separate from the log: (1) any significant factual error: misquoted pricing, wrong feature claims, misattributed statements; (2) any substantive negative coverage. One message each, with the quote and the link. Everything else goes in the sheet silently."
• "Do not reply, comment, or contact any author. Log and flag only."

Only two things get flagged

Ping the channel for every listicle inclusion and the team mutes it by week two, and then the misquote arrives into a muted channel. Errors and negative coverage are the two categories where response time changes the outcome: an error corrected on publication day is a friendly author email, while the same error at quarter-end has been cited by three other articles.

Reading the surrounding context is what separates this from keyword alerting. "Unlike [brand], which requires..." matches the same keywords as praise, so the sentiment column only means something because the paragraph got read.

Why log listicles and roundups at all?

They shape buying decisions and feed answer engines. Being absent from four "best [category] tools" roundups names exactly which authors to build relationships with, and being present-but-mischaracterized ("best for hobbyists" when you sell enterprise) is positioning drift you can correct at the source. The comparison category matters double if competitors run the same monitoring on you.

The reach-proxy column keeps the log honest about magnitude. A misquote in a 200-subscriber newsletter and one in a 90,000-subscriber one are different emergencies, and "unknown" is the honest cell for most podcasts.

The log becomes the coverage record

The quarterly coverage report assembles in minutes: mentions by month, outlet and type, sentiment trend, errors caught and corrected, the launch-week spike or its absence.

It is also the earned-media half of your answer-engine visibility, since coverage is most of what those engines read. Pair it with competitor monitoring for both directions. The no-reply rule stays permanent: a companion finds and files, and the comms person decides who hears from them.

Frequently asked questions