Alert rules
Create, edit, pause and delete the rules that decide which posts match and where they're sent.
An alert rule describes what to look for, where to look and where to send matches. Manage your rules under Rules in the sidebar.
Create a rule
Click New rule. The form has three steps; you can move between them freely and save once every required field is complete.
Rule: what should we look for?
- Prompt (required, up to 300 characters): a plain-language description of a matching post. AI checks every new post from your sources against it. See Writing good prompts.
- Name (optional): shown in your rules list. If you leave it blank, the prompt is used.
- Keywords (optional): a pre-filter applied before AI. Type a keyword and press Enter, or click Suggest with AI to get suggestions from your prompt.
- Classification (optional): tags and sentiment analysis that AI adds to every match.
- Status: active or paused.
Sources: where should we look?
Pick at least one source: a Facebook group, subreddit, review site, app store listing, blog or forum. You can add new ones by URL right here. Sources that are still in review or awaiting access can be selected too; the rule includes them once they're ready. See Sources.
Alert channels: where should matches go?
Pick at least one alert channel. Every match is delivered to each selected channel. If you don't have a channel yet, the form offers to create one.
Click Create rule. The rule starts matching new posts straight away.
Start from a template
Click Browse templates on the Rule step to start from a proven prompt. Templates are grouped by category: Lead generation, Competitor tracking, Brand monitoring, Complaints, Market research, Recruiting and Community. Fill in the blanks (empty fields use the example value) and click Use template.
Keywords
Keywords decide which posts are worth sending to AI, so they're your main lever on noise.
- Prefer phrases over single words:
project management toolbeatstool. - BuzzGlean rejects keywords with no letters, ones that are too short, and ones that appear in almost every post.
- When a keyword matches a large share of posts in your sources, you'll see a warning like "matches about 120 posts a day (35% of posts in your sources)". Consider a more specific phrase.
- Keywords count toward your plan as unique keywords across all rules: using
CRMin three rules counts once.
Tags and sentiment
Turn these on to label matches so you can filter and triage them in the feed:
- Tags: pick default tags or create your own. Default tags include Buy intent, Complaint, Competitor mention, Feature request, Bug report, Question, Praise and Churn risk. AI assigns up to three tags per match.
- Sentiment analysis: AI rates how the author feels about the brand, product or service: positive, neutral or negative. When it is off, new matches get no sentiment label (they still get tags) and don't appear under a sentiment filter in the feed. Turning it on applies to posts matched from then on.
Edit a rule
Open a rule from the list to edit it. Jump to any step, change what you need and save.
Pause or resume
Set Status to paused to stop a rule without losing its settings. Paused rules don't match posts or send alerts. Switch it back to active to resume.
Delete a rule
From the rules list, open the row's actions menu and choose Delete rule. The rule stops matching and sending alerts immediately. This can't be undone.
Rules that need attention
When a rule has sources that won't be monitored (Rejected or Can't access), the rules list shows a Needs attention banner and flags the rule. Use the Needs attention filter to see them all, open a flagged rule and click Remove them on its Sources step, or replace the sources. Sources that are only waiting (In review or Awaiting access) don't need any action. See Source statuses.
The dashboard also lists rules with no matches in 7 days, which may need a broader prompt or different sources.
Learn from feedback
Marking matches as good or not good in the feed helps BuzzGlean learn what you want from each rule. See Feedback.