Juma finds the earned coverage on the open web, scores every article against a fixed rubric with a verbatim journalist quote as the evidence, and returns a coverage audit plus an editable tracker.
Give Juma the client and their URL. It searches the open web for the coverage itself, opens each article, and separates reported editorial coverage from the investor commentary, aggregator recaps and wire reposts that inflate a clip count without telling anyone anything.
What comes back is a branded PR Coverage Quality Audit and an editable CSV tracker. Every scored article carries its points per criterion and one verbatim sentence from the journalist as the evidence, so a score can be argued with rather than just accepted. The audit ends with a next-quarter plan that separates a pitch an outlet could commission from a pre-brief ahead of a moment the outlet will cover anyway. 400+ marketing teams use Juma, and strategy, taste and judgment stay human: the audit gives the comms lead something to defend in the room, not a verdict to forward unread.
Coverage arrives one article at a time, and nobody wants to re-run a full audit to log a single hit. Paste the URL into the same chat and ask for one scored row in the existing format. Juma opens the article, applies the same fixed rubric, pulls the verbatim evidence quote, records whether an owned-domain link appears, and returns the row ready to paste into the tracker. Because the weights were fixed in the original audit, the new score sits on the same scale as everything already logged. Over a quarter this is what turns a pile of clips into a trend a comms lead can show a board.
Score this new placement into the same tracker, using the same fixed rubric and the same columns: [article URL]. Return one CSV row plus the verbatim evidence quote.
A placement can be prominent, positive and still miss the point, because the message the team briefed never made it into the journalist's own words. Give Juma the messaging framework and ask it to trace each message through the scored coverage. It reports which messages appear verbatim or closely paraphrased, which appear only in a quote from the spokesperson, and which never survived the edit at all. The output names the messages that travel on their own and the ones that only ever appear when someone from the company says them, which is usually the more useful half of the answer.
Using our messaging framework, trace each key message through the scored coverage: which ones landed in the journalist's own words, which appeared only inside a spokesperson quote, and which never made the edit.
Volume comparisons flatter whoever issues the most press releases. Quality comparisons are harder to argue with. Ask Juma to run the same audit on a named competitor across the same window and the same rubric, then put the two side by side. What comes back is the average quality score for each, where each brand is strongest, which outlets cover one and not the other, and which messages the competitor gets into print that the client does not. That last column tends to be the one that changes what the team pitches next quarter.
Run the same audit on [competitor] across the same six months and the same rubric, then compare: average quality, outlet overlap, and which messages they get into print that we do not.
An audit that stops at scoring tells a team what happened and leaves them to work out what to do. Ask Juma to carry the findings forward into a prioritised plan. Each row names the outlet, the angle, what the team has to supply to make it real, and whether it is a pitch or a pre-brief. That distinction matters: a business desk will not commission a brand's strategic framing, but it will take a concise fact sheet before results are announced. Separating the two stops a plan promising placements no editor was ever going to run.
Turn the audit into a prioritised next-quarter plan. For each row give the outlet, the angle, what we need to supply, and label it PITCH or PRE-BRIEF.
One Project per client. Juma remembers the brand's messages, audiences and past coverage without re-briefing, and every flow run in that project starts from the same foundation. For a coverage audit this changes the two criteria that are otherwise guesswork: message pull-through is scored against the messages the team actually briefed, and audience fit is scored against the segments the client actually sells to.
Messaging framework
The key messages, proof points and positioning the team briefs into every story. With this in the project, message pull-through stops being an impression and becomes a check against the real list.
Audience and segment definitions
Who the brand sells to, in the client's own words. Audience fit is the criterion most often scored on assumption, and this is what grounds it.
Previous coverage audits and trackers
Last quarter's scored tracker. With it in the project, the new audit opens with the trend rather than starting the method argument again.
Spokespeople and available proof
Who can speak, on what, and which data the team can hand a reporter. The pitch plan then asks for things the client can actually supply.
Each knowledge item takes a one-line description telling Juma how to use it. Keep them plain and specific.
Juma searches the open web for articles about the brand, opens each one, and scores it against a fixed rubric the team sets in the prompt: outlet authority and reach, key message pull-through, prominence, sentiment and framing, and audience fit. Each article gets its points per criterion and a total out of 100.
The part that makes the score usable is the evidence. For message pull-through, Juma quotes one verbatim sentence from the journalist's own text, so the score points at something specific in the article rather than asserting a judgment. Anything that cannot be observed from a public page stays out of the points entirely.
Because an unpinned rubric gets reinvented every time. While building this flow we ran the same audit three ways without fixing the weights, and one TechCrunch feature came back at 95, 67 and 91 out of 100. The article was identical each time. The criteria and their maximum points were not.
Naming the criteria and their point values in the prompt fixes this, and it is why the saved flow carries the rubric in its own instructions. Scores then sit on one scale across articles, across quarters, and across the people running the audit.
Scored coverage is editorial: reported articles written by a journalist about the brand. That covers product and feature stories, corporate and executive coverage, campaign and marketing coverage, reported business stories, reviews and features. The test is whether a journalist wrote it about the brand, not whether the brand's name appears in it.
Investor and stock commentary, course and app aggregator recaps, press release reprints, company investor-relations announcements and syndicated reposts of a wire story already scored once are excluded. They appear in their own table with the reason, so nothing disappears silently and the client can see the call that was made.
No, and it is not trying to. A monitoring platform watches continuously and catches far more than a search of publicly findable coverage will. This flow does the part the platform leaves to a human: judging whether a placement was any good and what to do next.
The audit states its own limits on the page. It is a sample rather than a full clip set, the scores are output measures rather than business outcomes, and it never estimates advertising value equivalency. Teams running a monitoring tool usually paste its export in and use this for the scoring and the plan. Human review on every output.
Two files. A branded PR Coverage Quality Audit PDF that opens with three instructions for the quarter, then the scored breakdown, the exclusions, what the pattern says about which messages land and which drift, and a prioritised plan separating pitches from pre-briefs. And an editable CSV tracker with one row per article and one column per criterion.
The tracker is the part that keeps working. Score each new placement into the same columns and the quarterly review becomes a trend line instead of a fresh debate about method.
Name a brand and its competitors, and this share of voice analysis returns two numbers: the share of Google visibility the brand owns, and the share of AI answers that name it.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
.webp){kind=icon}
