Automations

An automation rule pairs conditions with actions: when morale drops below 60, then send a high-urgency alert. Automations live at /signals/automations.

Every rule has the same anatomy — a name, one or more conditions combined with AND or OR, one or more actions, an optional team scope, and a delivery channel. There are three ways to get one.

Start from a template

Thirteen pre-built templates cover the common cases. This is the fastest way to a working rule and the best way to see how conditions are meant to be shaped.

1. Scroll to Template Library and filter by category — Team Health, Performance, Engagement, or Operations.
2. Select a template. The rule form opens pre-filled with its conditions and actions.
3. Adjust the thresholds to your numbers. The templates ship with plausible defaults, not your defaults.
4. Save.

The templates:

Template	Category	Condition	Action
Morale Drop Alert	Team Health	Morale Score less than 60	Send Alert (high)
Burnout Early Warning	Team Health	Burnout Risk greater than 75	Send Alert (high) + Schedule 1:1
Team Health Report	Team Health	Morale Score less than 70 or Burnout Risk greater than 60	Generate Report (team health)
Psychological Safety Watch	Team Health	Psychological Safety less than 55	Send Alert (medium) + Create Action Item
Velocity Slowdown	Performance	Deployment Confidence less than 40 and Cycle Time greater than 5	Generate Report (trend analysis)
Deploy Confidence Drop	Performance	Deployment Confidence less than 50	Generate Report (team health) + Send Alert (medium)
Sprint Health Check	Performance	Deployment Confidence changes by 15 over 14 days	Generate Report (signal summary)
Engagement Decline	Engagement	Engagement less than 50	Create Summary (team) + Flag for Review
Feedback Silence	Engagement	Feedback volume less than 10	Create Summary (team) + Send Alert (medium)
Sentiment Shift Alert	Engagement	Morale Score changes by 20 over 7 days	Send Alert (high) + Create Action Item
PR Backlog Warning	Operations	PR Count greater than 20	Send Alert (medium) + Flag for Review
High Issue Volume	Operations	Issue Count greater than 30	Flag for Review + Send Alert (medium)
Weekly Signal Digest	Operations	Engagement at most 100	Create Summary (organization)

Describe it in plain English

1. In Create a rule with AI, type what you want — for example, Alert me on Slack when any team's morale drops below 60%. At least five characters.
2. Select Generate. The rule form opens and the description is parsed automatically.
3. Review the conditions and actions the AI produced, adjust anything it got wrong, and save.

Gibberish gets refused rather than turned into a rule you didn’t mean.

Build one from scratch

Select Create Rule and fill in the form directly.

1. Name — up to 200 characters — and an optional description, up to 1,000.
2. Under When, add conditions and pick AND or OR to combine them. The logic applies to the whole set, not per pair.
3. Under Then, add actions.
4. Set Team Scope — a specific team, or All teams.
5. Set Delivery Channel — Default, Slack, or Email.
6. Save.

Condition reference

Each condition is a type, a field, an operator, and a value.

Condition type	Use it for
Signal Threshold	A signal sitting above or below a level
Sentiment Change	Sentiment moving over a window
Feedback Volume	How much feedback is arriving
Trend Direction	A metric trending over a window
Integration Event	A metric from a connected source

Signal, sentiment, volume, and trend conditions choose from Morale Score, Burnout Risk, Engagement, Psychological Safety, and Deployment Confidence. Integration conditions choose from PR Count, Issue Count, Deploy Frequency, and Cycle Time — these come from your connected sources, so a rule on Cycle Time needs the relevant integration in place.

Operators are less than, greater than, equals, at most, at least, and changes by.

Sentiment Change and Trend Direction take a time window in days, between 1 and 365. The other three types don’t — they read the current value.

Note that the automation fields are a shorter list than the signal catalog. Conditions can be built on these five signal fields, not on all fifteen rules.

Action reference

Action	Configuration
Generate Report	Report type: Team Health, Trend Analysis, or Signal Summary
Send Alert	Message template, and urgency: Low, Medium, or High
Create Summary	Scope: Team or Organization
Schedule 1:1	None
Flag for Review	None
Create Action Item	None

A rule needs at least one condition and at least one action.

Managing rules

New rules start as Draft. The toggle on each rule card switches between Active and Paused; the status filter above the list narrows to Active, Paused, or Draft, and search matches on name.

Each card reports how many conditions and actions the rule has, how many times it has been triggered, and when it last triggered.

Caution

Check Triggered nx and Last triggered on the card after you activate a rule. Those two fields are the only confirmation that a rule is doing anything — a rule can be perfectly well-formed, Active, and simply never meet its conditions.

The overflow menu on each card offers Edit and Archive.

Practical guidance

Set thresholds from your own numbers. Open Signal Intelligence first and look at what your teams actually score. A morale threshold of 60 is meaningless if your teams never drop below 75 — and deafening if they hover at 55.

Scope to a team while you’re tuning. All teams on an untested rule multiplies whatever you got wrong by the number of teams you have.

Combine conditions with AND to cut false alarms. The Velocity Slowdown template is the pattern to copy: a signal condition and an integration condition together, so a single noisy metric can’t trigger it alone.

Use changes by rather than an absolute value for anything you’re watching for movement. A team that has always sat at 55 isn’t news; a team that dropped 20 points in a week is.

Related

• Signals overview — where the signal values come from
• Signal events and evidence — actions, and how they differ from automations