buzzabout docsAnalyse feature requests
Apply the Feature requests skill to a collected dataset to turn social chatter into a severity-ranked, triaged product backlog.
If you build a B2B SaaS project-management tool, your roadmap signals are already sitting in public conversations — people asking "please add a Gantt view", "why doesn't it support SAML SSO", "I wish it integrated with Slack". The Feature requests skill reads a dataset you've already collected, pattern-matches those product-directed asks, ranks them by severity, and triages each one, so a stream of social chatter becomes a prioritised product backlog you can hand to your product team.
Prerequisites
- A dataset with a completed collection run, so there are posts and comments to analyse. See Set up your project and Create a mention collection.
- A chat or project open against that dataset — a skill can only be applied inside one. If you're new to applying skills, start with Use skills.
Walkthrough
Open skills
Open a chat scoped to the dataset that holds your project-management mentions. In the chat composer, click the Open skills book icon to pop the skills picker. The skill needs a draft to attach to, so make sure you're inside a chat or project.
Browse all skills
The picker opens on Favourite skills. Click Browse all skills in the footer to open the full library, then open the Content & positioning or Audience research category in the left Skills sidebar — Feature requests is tagged under both. You can also type into the search box to jump straight to it by name.
Use this skill
Click the Feature requests row to open its detail view and read the instructions. When it's the right one, click Use this skill to attach it to the current chat draft. The modal closes and the skill appears as a chip in the composer.
Add context and send
Attach the project-management dataset as context in the composer before you send — the skill chip and the context chip sit independently on the same message. Type a short prompt such as What features are people asking us to add? and send. The skill's instructions tell the assistant to scan for product-directed asks, so you don't need to spell out the method.
Let the pattern pipeline run
To separate genuine feature requests from general complaints, the skill runs the async pattern-matching pipeline over your mentions. This takes a few minutes and may charge credits. If the dataset already has pattern assignments from an earlier run, the assistant reuses them instead of re-detecting, so a second pass is faster and cheaper.
The pipeline pattern-matches asks like "please add X" and "why doesn't it support Y" and deliberately keeps them distinct from pain points (problems with what already exists) and objections (reasons not to buy). It ranks each request by severity and triages it: requests that are already solved are flagged as a discoverability gap rather than a build, and the rest are scoped small, medium, or large.
Read the severity bar and request table
The assistant returns the standard skill artifact: a hero chart, a detail table, and three takeaways. The hero chart is a severity bar colour-coded by recommended action — ship, communicate (for already-solved asks that just need surfacing), or defer. Below it, a request detail table lists each request with its severity, scope, and triage call, and the takeaways name the top 3 to prioritise. Open the result in the side panel to read the full breakdown and the mentions behind each row.
Gotchas
- The pattern pipeline is asynchronous and may charge credits. It polls for several minutes. The assistant reuses existing pattern assignments on the dataset when they're present, so it only pays the full cost on the first run; re-running over the same data is faster.
- Feature requests are not pain points or objections. The skill deliberately scopes to product-directed asks ("add X", "support Y"). Problems with the existing product are pain points; reasons people don't buy are objections. If you want those, use the matching skills instead — see Next steps.
- Already-solved asks are a discoverability signal, not a build. When people request something you already ship, the skill triages it as
communicaterather thanship— the work is surfacing the feature, not building it. - You need collected mentions first. Pattern matching requires a dataset that already has posts and comments — run a collection before applying the skill.
- Applying the skill does not run it. Attaching the chip only pins the instructions to the draft. The analysis (and any credit cost) happens when you send the message.
Next steps
Analyse pain points
Turn a collected dataset into a ranked map of what your buyers actually complain about, using the built-in Analyze pain points skill.
Track share of voice
Benchmark your brand against competitors by share of the conversation, with per-brand sentiment, using the Mentioned brands skill or an assistant ask.