The AI Triage Engine automatically routes referrals to the correct branch and area, checks carer capacity, and applies automation rules to accept, decline, or flag referrals for manual review. Branch Managers can configure the AI Triage Engine independently for each branch.

Triage uses a two-step process:

Three automation rule types are available, listed here in priority order:

Rules are configured at branch level by Branch Managers. Each rule can combine multiple conditions using AND logic. To achieve OR logic, create separate rules for each condition.

Conditions available within rules include: branch, area, postcode, age, referrer organisation, individual referrer, and carer capacity (both standard and preferred).

Trial Mode lets you test automation rules safely before activating them. When a rule in Trial Mode would fire, the system shows a recommendation to the coordinator instead of auto-actioning it.

The coordinator can accept or decline the suggestion, selecting a justification reason if they decline. Performance statistics, times applied, accepted, declined, and acceptance rate, are visible per rule.

Once you're confident a rule is working correctly, turn off Trial Mode to activate full automation.

To re-run triage on an existing referral:

This re-runs Branch/Area mapping (only if not already set), recalculates capacity, and re-applies automation rules. This is useful if rostering data has changed or if you need to reassess a referral.

In the AI Triage Engine under Availability Configuration, you select which availability and unavailability types to include in the capacity calculation. This is configured at branch level.

This is important because short-term unavailability for example, a cold, may be irrelevant for a six-month care package assessment, while long-term unavailability for example, regular child commitments, should be included.