buzzabout docsGenerate a content brief
Run your analysis in one chat thread — pain points, feature requests, content ideas, a content gap analysis — then ask the assistant for a content brief that pulls it all together into a shareable publishing plan.
A content brief isn't a button — it's an asset the assistant generates inside a chat, and you can produce one in any chat, not just from a project folder. The trick is that the brief is the culmination of the analysis you've already done in the same thread. Say you run an e-bike DTC brand: in one chat grounded on your Reddit and TikTok mentions about commuting, range anxiety, and theft, you first surface pain points and feature requests, then get content ideas, then run a content gap analysis of your niche. Ask for the content brief last and it weaves all of that in — an effort-vs-impact plan where every angle traces back to a real finding you already pulled up. The result is a shareable asset you can hand to teammates.
Prerequisites
- A mentions dataset with a completed collection run, so there are real posts to ground the brief on. See Create a mention view.
- Enough volume for the synthesis to find repeatable angles — a few dozen mentions surface anecdotes; aim for ~100+ before you trust the brief.
Walkthrough
The order matters here. Each skill below builds on the last in the same chat thread, so by the time you ask for the brief, the assistant already has pain points, requests, ideas, and gaps in front of it.
Open a chat
Start a chat from the home page, or open any project's chat — either works. The brief is just an asset the assistant emits, so you don't have to begin from a project folder.
Add context
Ground the chat in a real dataset. Click Add context in the composer and tick a mentions dataset — under In this chat, All your mentions datasets, or In this project if you're in a project. For the e-bike brand, pick the dataset holding your commuting and range-anxiety threads.
If you have nothing to attach yet, you'll see No mentions datasets yet. or No mentions datasets in this project. — collect mentions first, then come back. The brief is only as good as the conversation underneath it.
Analyse pain points and feature requests
Apply the Analyze pain points skill, then the Feature requests skill, in the same thread. Open the skills picker (the book icon, Open skills), click Browse all skills, find each under Content & positioning, and choose Use this skill.
Pain points surface the recurring frustrations — range anxiety, theft worry — that make the most magnetic hooks. Feature requests turn product-directed asks into explainer angles. Running these first means the assistant has the raw frustrations and asks already on the table before it ever drafts a brief.
Get content ideas
Now apply the Suggest content ideas skill in the same chat. Because it's a synthesis layer, it draws on the pain points and requests you've already surfaced — plus outliers, audience questions, and buzzwords — and pulls candidate ideas, each traced to a real finding, scored on effort versus impact.
Add a short steer if you want, for example Focus on short-form video for commuters, and send.
Run a content gap analysis
Apply the Content gap analysis skill next. It compares what your niche talks about (supply) against what your audience is asking for in comments (demand) and surfaces a demand-vs-supply quadrant — the gaps where demand outstrips supply are your sharpest, least-crowded angles.
Running this in the same thread means the brief that follows knows which ideas are genuinely under-served, not just popular.
Ask for the content brief
Now ask, in plain language: Pull all of this into a content brief. Because everything above lives in the same thread, the assistant incorporates the pain points, feature requests, ideas, and gaps you already surfaced — it doesn't start from scratch.
It composes the brief as an asset: an effort-vs-impact scatter (your quick wins sit high-impact / low-effort), followed by a content-brief table — one row per idea, each carrying a hook, a structure, a format, an emotion, and a CTA, plus a source quote pulled from your mentions. In the thread the asset shows as a compact reference card with an Open in side panel action; when streaming finishes, the side panel auto-opens with the full brief.
Draft the script
At the bottom of each brief row there are follow-up buttons — by default Draft the script, Find similar mentions, and Suggest a hook. Each button's label is the exact message sent back to the assistant to continue from that specific idea, so clicking Draft the script on your "range anxiety, debunked" row asks the assistant to draft that script next.
Share it with your team
The brief is an asset, so you can hand it off. From the side panel, use Share asset to get a public Sharing link or an Embed snippet to drop into Notion or a blog, or download it as PNG or PDF. See Share & export assets for the full set of controls — and note that share links are public and unauthenticated.
Gotchas
- It's a use case, not a button. There is no "Generate content brief" control anywhere. You ground a dataset, run your analysis in the thread, then ask — and the model decides to emit the brief asset. A vague prompt with no grounding may not produce one.
- Order matters. The brief is the culmination of the chat, not the start of it. If you ask for a brief cold, before surfacing pain points, ideas, and gaps, you get a thinner result — the synthesis only weaves in what's already in the thread.
- Ground it in a real dataset. Data-backed briefs need a mentions dataset attached via Add context. The picker is datasets-only; wherever product copy says "collection" it means a mentions dataset.
- The skills trigger underlying passes. Pain points, content gap, and idea synthesis kick off pattern-detection runs. Those run asynchronously, can take a few minutes, and may charge credits; if the data is too thin the assistant will say so rather than invent angles.
- The inline card doesn't show the body. In the thread you only see the reference card; the full brief parses in the side panel. The panel auto-opens only for an asset that finished streaming this turn — reopening an old chat won't auto-open historical briefs, so click the card.
- Web-only rendering. Over MCP,
buzzabout__askreturns markdown only — the rich brief asset, its scatter, and the follow-up buttons exist solely in the web app.