IT incidents — outages, degraded services, access issues — are the most frequent ITSM use case. Without a structured workflow, they pile up in email inboxes, spreadsheets and chat threads, generating SLA breaches, duplicate tickets and frustrated users.

Utilyx models incident management as an executable workflow with AI-powered triage, automatic priority assignment, SLA tracking and resolution suggestions from the knowledge base.

The workflow is structured as a decision tree with 3 main branches (P1/P2, P3, P4) and 8 steps end-to-end, plus an L2 escalation sub-branch. Every step is executed by a named actor (User, AI Agent, Routing engine, Technician, L2 support, IT Manager) and every action is timestamped in the audit trail.

• Step 1 (Start — User): the user reports the incident via the self-service portal, email or chat. Mandatory fields: title, description, affected service, urgency, attachment (screenshot/log). The form opens the conditional sections dynamically based on the urgency selected.

• Step 2 (AI Agent): the Copilot IA classifies the incident (incident, request, problem), assigns priority P1-P4 based on impact × urgency matrix, and deduplicates against open tickets by matching title, affected CI and user within a 24h sliding window.

• Step 3 (Routing): three conditional branches split the tree — Branch A (P1/P2): immediate escalation with email to IT manager, Slack notification via call-connector and SLA timer 4h/8h; Branch B (P3): standard investigation with KB article suggestions from the AI and technician diagnoses; Branch C (P4): self-service where the AI chatbot resolves from the KB or escalates to an agent if confidence < threshold.

• Step 4A (P1/P2 path): investigation with KB suggestions + CMDB lookup via sql-query (SQL query on affected CIs to map dependencies and blast radius). The technician documents diagnosis in the timeline_incident subtable.

• Step 5A: resolution is attempted; if MTTR exceeds the SLA threshold, the sub-branch L2-escalation fires — the ticket is reassigned to L2 support, the IT manager is notified and the SLA clock pauses during the handover.

• Step 6A (usePDF): an incident report is auto-generated from a Word template with merge fields (timeline, diagnostic, resolution, RFO) and converted to PDF for signature.

• Step 7A (useArchive): legal archiving of the incident report + attachments with index fields incident_id, date, severity, category, ci_affected — encrypted and tamper-evident for the legal retention period.

• Step 8A (Closure): the user confirms resolution, a KB article is auto-generated from the resolution text, and an NPS survey is sent. Detractor scores reopen the case for managerial review.

Teams that adopt Utilyx report MTTR reduced by 40% thanks to AI-assisted triage and resolution suggestions, 60% of P4 tickets auto-resolved by the AI chatbot without human touch, and zero SLA breaches with predictive alerts that fire 2h before the deadline.

The 7-language KB search lets a single support team serve a multilingual user base, the visual routing designer lets any IT analyst re-prioritize branches without code, and the no-code workflow builder turns a plain-language description into a runnable ITSM process — 60% faster deployment, 70% lower TCO, no consultants required.