buzzabout docsSpot trends
Apply the What's trending skill to a collected dataset to catch rising topics early and ride them this week.
If you run a DTC e-bike brand, the difference between a good quarter and a flat one is often timing — catching a topic like torque-sensor upgrades, commuter-range anxiety, or a viral fold-and-carry format while it is still rising, not after every competitor has posted about it. The What's trending skill reads a dataset you've already collected and tells you what is genuinely accelerating: it defines a trend as change over time, so it splits your mentions into a recent window versus a baseline window, compares topic clusters on volume, reach, and cross-creator adoption, and shortlists the clusters that are actually rising — not a single mega-post that briefly spiked. The output is a multi-line over-time chart, a trend detail table, and three plays you can ship this week.
Prerequisites
- A dataset with a completed collection run, so there are mentions to compare across time. See Set up your project and Create a mention collection.
- Timestamped mentions and enough volume. A trend is change over time, so the skill needs posts with usable timestamps spread across a recent and a baseline window. A thin or single-day dataset has nothing to compare — collect more before running it.
Walkthrough
Open skills
Open a chat against the dataset that holds your e-bike mentions — a skill needs a draft context to attach to. In the chat composer, click the Open skills book icon, or type / at the start of an empty message to pop the skills picker.
Browse all skills
The picker shows your Favourite skills with a Browse all skills footer. Click Browse all skills to open the full library.
Market research
In the modal, open the Market research category in the left Skills sidebar — What's trending lives there. You can also type What's trending into Search skills… at the top to jump straight to it.
Use this skill
Click the What's trending row to open its detail view and read the instructions — it explains that a trend means change over time and that it shortlists multi-creator clusters rather than one-off spikes. When you're ready, click Use this skill to attach it to the current chat draft. The modal closes and the skill appears as an inline chip in the composer.
Add context and send
Attach the e-bike dataset as context via the composer's Contextualize picker, type a short question such as What is trending in the e-bike conversation right now?, and send. The skill instructions tell the assistant to split the mentions into a recent window versus a baseline, compare topic clusters on volume, reach, and cross-creator adoption, and keep only the clusters that are genuinely rising.
Read the over-time chart and trend table
The assistant returns its deliverable as an Asset. At the top is a multi-line over-time chart — one line per shortlisted cluster — so you can see which topics are pulling away from the baseline. Below it is a trend detail table giving each cluster its magnitude (how big the change is), its catalyst (what's driving it), and how to ride it. Read the lines that are climbing steepest and cross-check them against the table before you act.
Ship the three plays
The skill closes with three plays to ship this week — concrete, time-boxed moves tied to the rising clusters, such as a post format to test, an angle to lean into, or a question to answer while the topic is hot. Pick the ones that fit your roadmap and act on them before the window closes; the value of a trend is in catching it early.
Gotchas
- A trend is change over time — you need timestamps and volume. The skill compares a recent window against a baseline, so mentions must carry usable timestamps spread across both. A thin, single-day, or low-volume dataset has nothing to compare against and the skill will say so instead of inventing a trend.
- Single-post surges are flagged, not celebrated. A cluster that spikes off one mega-post is not a trend. The skill shortlists clusters with cross-creator adoption (multiple creators, not one viral post) and explicitly flags single-post surges so you don't chase a fluke.
- Watch for dated catalysts. The skill also flags catalysts that have already passed. If the thing driving a cluster is an event that's over, the "trend" may already be cooling — read the catalyst column before you commit a play.
- It runs a pattern pipeline that's asynchronous and may charge credits. Comparing topic clusters runs a pattern-matching pipeline that can take a few minutes; if a pattern already exists on the dataset it is reused rather than re-run.
- It's a point-in-time read. Running the skill once tells you what's rising today. To keep catching new trends as they emerge, pair it with a listening agent that re-runs the underlying queries on a schedule so the dataset stays fresh — then re-run the skill against the growing data.
Next steps
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.
Generate 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.