Signals
A signal is something DigitalCore noticed in your data and thinks you should look at. Every signal can be turned into a decision case with one click — the case will carry the signal as the AI agent input and a scenario will be generated automatically.
What each signal card shows
| Field | Meaning |
|---|---|
| Type | The kind of issue (margin pressure, cost escalation, SLA breach risk, and so on). |
| Severity | Critical, High, Medium, or Low. Set in Thresholds. |
| Confidence | A 0 to 1 score of how clear the pattern is. |
| Engagement | Which engagement is affected, or “portfolio” for cross-engagement issues. |
| Recommended playbook | If the issue matches a known pattern, the suggested response. |
| Respond | Opens the Response screen so you can act on it. |
Common signal types
| Signal | Raised when |
|---|---|
| Margin pressure | Engagement margin is trending toward Amber or Red. |
| Cost escalation | Actual costs exceed plan by your set percentage. |
| SLA breach risk | A KPI is trending toward a breach within your set window. |
| Capacity overrun | Recorded hours are above plan. |
| AI budget overrun | AI usage cost is above the warning level. |
| Health degradation | Engagement health score moved to a worse colour. |
| Correlation | A pattern crossing two areas, for example SLA breach plus a capacity shortage. |
| Outcome miss | Actuals are clearly different from the plan. |
Full list in Concepts: Signals.
Filter and sort
Filter by severity, engagement, signal type, source, or status (open, responded, resolved). The default order is severity first, then most recent.
Acting on a signal
Click Respond to open the Response screen. DigitalCore will:
- Generate two or three options for handling the issue.
- List the assumptions behind each option.
- Wait for you to approve and apply, or decline.
Applied responses are kept in History.