Resolving Alerts

The Impact Cascade Engine doesn’t just tell you what broke — it tells you how to fix it with specific suggested actions.

Suggested Actions

Every cascade alert includes one or more suggested fixes. These are specific, actionable recommendations:

Examples of Suggested Actions

AlertSuggested Action
”Task duration changed to 90min""Extend these 8 time blocks from 60min to 90min"
"User Ana deactivated""Reassign Email Triage to David (trained, available)"
"SOP updated to v3""Notify 5 team members about updated procedures"
"Workflow step assignee missing""Assign Kevin as replacement for step 3”

Resolving an Alert

One-Click Resolution

For straightforward fixes:

  1. Open the cascade alert
  2. Review the suggested action
  3. Click “Apply” or “Accept” to execute the fix
  4. The alert is resolved and changes are applied

Manual Resolution

For complex situations requiring judgment:

  1. Open the cascade alert
  2. Review the impact chain
  3. Take your own corrective action (reassign tasks, update schedules, etc.)
  4. Click “Mark Resolved” to close the alert

Dismissing Alerts

For alerts that don’t require action:

  1. Open the alert
  2. Click “Dismiss” with an optional reason
  3. The alert is removed from your active list

The Impact Summary View

For large cascades affecting many entities, use the Impact Summary View:

  • Severity breakdown — how many critical, warning, and info alerts
  • Entity breakdown — how many tasks, time blocks, users affected
  • Batch resolution — apply suggested fixes to multiple alerts at once
  • Audit trail — see who resolved what and when

Best Practices

  1. Prioritize by severity — critical first, then warnings, then info
  2. Review before applying — suggested actions are smart but verify they make sense for your context
  3. Use batch resolution when multiple alerts have the same fix
  4. Check the audit trail to ensure nothing was missed

The goal: When you make a change in BFF, you should be able to see and resolve all downstream effects within minutes — not discover broken processes days later.