How does the Activity View get its insights?

Classification Signals

Each classified item displays the specific signals that triggered its classification — shown as discrete labels with supporting data. These signals are generated instantly using a template-based engine, no AI involved.

Example signals by category:

Needs Attention:

• "Dependency at risk"
• "Confidence degraded: Help needed"
• "Overdue by 12 days"
• "Progress regressed: -15%"

Not Updated:

• "No check-ins for 14 days"
• "No recent activity"

Progressing Well:

• "Confidence improved: All good"
• "Progress improved: +10%"
• "Recent activity: 2 days ago"
• "3 check-ins this period"

Each item shows all applicable signals, making it clear at a glance what's driving the classification.

Dependency Intelligence

Activity View fetches and classifies every connected dependency for each work item — upstream blockers, downstream consumers, delivering items, and child objectives. When a dependency is degraded, the parent item is flagged automatically.

This means:

• Blocked-by detection — A key result depending on an unresolved interlock is flagged as "Needs Attention" with a dependency risk signal, even if the key result itself looks healthy.
• Child-to-parent bubbling — An objective is flagged when its child key results are struggling or have low confidence, so leadership sees problems at the strategic level without drilling into individual items.
• "Needed for" visibility — Items show what depends on them downstream, making the impact of delays visible before they cascade.
• Cross-team surfacing — When Team B's interlock regresses from resolved to unresolved, Team A's dependent key result is immediately flagged — without waiting for Team B to report the change.