Say you track 200 prompts and spend a month publishing content around one topic. If the work lands, that topic’s visibility should jump — but averaged across 200 prompts, a big win on 14 of them barely moves the headline number. Today BotScope gets tags: label your prompts, then segment every chart, report, and API call down to exactly the slice you care about.
TL;DR
- Every chart gets a segment picker — filter the whole pulse view, categories, competitors, providers, and sentiment to prompts carrying one tag.
- Your prompts are already tagged. Generated prompts carry their topic, category, and audience automatically — segmentation works today, with zero setup.
- Add your own campaign tags — label any set of prompts (one at a time or in bulk) with something like
spring-content-push, and measure that initiative on its own. - Tags are retroactive — tag a prompt now and your existing scan history segments instantly. No waiting for the next scan.
- Reports carry the segment — share links and PDF exports can be frozen on a segment, clearly captioned, so a client sees the campaign’s numbers.
- The API and MCP server speak tags too — ask Claude “how has Topic A performed this month?” and it queries just that slice.
Why we built it
AI visibility work is rarely spread evenly. You pick a front: a topic cluster to own, an audience to win, a quarter’s content campaign. Then you scan, and the dashboard answers a different question — how is the whole brand doing? — when what you asked was did the thing I just did work?
That gap matters most at exactly the moment you need confidence. An agency that shipped fifteen articles on one theme needs to show the client that those prompts moved. An in-house team choosing next quarter’s focus needs to know which past pushes actually paid off. The average can’t tell you; the segment can.
Tags turn your prompt library from one undifferentiated pool into a set of measurable fronts.
What you can segment by
The built-in axes — no setup. Every prompt BotScope generates already knows where it came from: the category it probes, the audience it speaks as, the topic it asks about. Those labels are on your prompts right now, which means the segment picker works the moment you open it.
Your own tags — for everything else. Campaigns don’t always map to one topic. So you can stamp any free-form tag on any set of prompts — spring-content-push, post-rebrand, priority-pages — and segment by that. This is the one you’ll reach for when the question is “did this initiative work?”
Tagging lives in the prompts editor (Settings → Tracked prompts): edit tags on a single prompt, or check a batch and tag them all in one click. A filter above the list shows you exactly which prompts carry which tag.
Where segments show up
Every visibility chart. The segment picker sits next to the date picker on the Pulse, Categories, Competitors, and Providers tabs — pick a segment and scores, trends, share of voice, and competitor rankings all recompute on that slice, in lockstep. Sentiment segments too. The selection lives in the URL, so a segmented view is a shareable link your teammates open to the same picture.
Client reports. When you create a share link or export a PDF with a segment active, the report is frozen on that slice and clearly captioned — “Segment — Topic: sustainable packaging” — so there’s no ambiguity about what a client is looking at. One click removes the segment if you want the full-account report instead.
The API and MCP server. Every data endpoint accepts category, useCase, topic, and tag parameters:
curl -s -H "Authorization: Bearer $BSK"
"$BASE/projects/$PID/visibility?tag=spring-content-push"And because the MCP tools take the same parameters, the assistant you’ve connected can answer segmented questions directly:
How has our sustainable packaging content push performed over the past month?
It's working. Across the 14 prompts tagged sustainable packaging, Acme's visibility score went from 18 to 58 over the month — against a company-wide average that only moved 5 points, so the lift is coming from this work specifically.
- L4 Citations 6 → 56 — the new guides are now cited directly by the models.
- L3 Recommendations 10 → 47 — Acme appears in most “best sustainable packaging” answers.
- Share of voice 31% on this topic, up 14 points — Globex still leads at 42%, but the gap halved.
The biggest remaining gap is Perplexity, where Globex's comparison page is cited and yours isn't yet. Want me to pull those specific responses?
FAQs
Do I have to tag anything before this is useful? No. Generated prompts already carry their topic, category, and audience — the segment picker is populated from day one. Custom tags are only needed for groupings BotScope can’t infer, like a campaign.
Are tags retroactive? Yes. Segmenting filters your existing scan history through each prompt’s tags as they are now, so a tag you add today reshapes charts of last month’s data instantly. The one exception: if a prompt was deleted and regenerated at some point, results from its earlier incarnation can’t be attributed to the new prompt — history from before a “Regenerate defaults” won’t appear under a segment.
Do tags change my scores or consume credits? No on both. Tags are metadata — scans, scores, and history are untouched, and segmenting is free. The unsegmented view is always one click away.
Can I combine tags? In the dashboard you segment by one tag at a time — deliberately, so a chart always has one unambiguous meaning. The API accepts multiple filters at once (a topic and a tag, or several values of each) for more surgical queries.
What happens if a segment has no data? You’ll see an honest empty state — zeros, not the unfiltered numbers quietly standing in. If a segment shows nothing, those prompts genuinely haven’t been scanned yet (or were regenerated since).
Does editing tags mark prompts as “manually edited”? No. Tagging never changes the prompt itself, so BotScope still manages generated prompts for you. One consequence worth knowing: Regenerate defaults recreates non-edited prompts from scratch, so custom tags on them are lost — the confirmation dialog warns you before that happens.
Is this available on my plan? The segment picker and tagging are available now on every project. The API and MCP tag parameters are included wherever API access is (Starter and up).
The teams that win in AI search don’t optimise “the brand” — they pick a front, do the work, and check whether it moved. Tags close that loop: label the work, run your scans, and let the segment tell you — in your dashboard, your client’s report, or a conversation with Claude — whether it paid off.