AI-assisted service desk support
A new team member needs reporting access. One clarification has gone nowhere. Compare the request with three authored suggestions for the Service Desk.
Fictional Harbor & Pine example and employees · Authored draft awaiting review · No live AI or recorded AI output. As of 20 November 2026, 14:00 UTC. People approve access; nothing is sent or fulfilled here.
Context comes from the request, service catalog, access definitions, local policy, attached approval and current knowledge article. View scenario sources.
The aim is a more complete request, clearer ownership and fewer avoidable clarification exchanges. Those are benefits to evaluate, not measured results. The request remains open: the employee clarifies the need, the approvers decide and Business Applications fulfills approved access.
AI could help interpret the reporting task, select relevant knowledge and draft a clear reply. Required-field checks, approval enforcement and routing can use ordinary software rules. Neither mechanism should copy a colleague’s access or promise a completion date without authority.
Business & operationsChoose request types and agree what staff must clarify, approve and fulfill.
Technical implementationMap ticket fields and approval states. Keep recommendations separate from actions.
Business & operationsAssign catalog and article owners. Keep role definitions and applicability current.
Technical implementationRetrieve permitted sources with version, effective date and audience. Expose missing or conflicting guidance.
Business & operationsDefine approval scope, fulfillment ownership and how exceptions are decided.
Technical implementationUse software checks for required fields, approval gates and routing. Restrict access to request and approval data.
Business & operationsReview suggested roles, questions and wording. Avoid requesting information already supplied.
Technical implementationProvide relevant request history and knowledge. Treat source text as data; return cited drafts, not access commands.
Business & operationsKeep the Service Desk, approvers and fulfillment team’s responsibilities clear.
Technical implementationShow drafts in the existing ticket. Retain original records and approval history; integrate only where needed.
Business & operationsReview relevance, missed approvals and unnecessary questions. Measure usefulness with staff before expansion.
Technical implementationTest wrong-role, stale-article and missing-approval cases. Monitor source freshness, quality and failures.
Start with your service catalog, knowledge and approval workflow. Existing features can support parts of this example; local rules and source access still need preparation.
Which guidance applies, and what information should we gather?
The virtual service agent’s AI answers summarizes linked knowledge with source articles. Configured intents can gather information before a human handoff.
Atlassian documentation ↗Maintain applicable articles and configure the request flow. Preserve approval gates; a knowledge answer does not authorize an entitlement.
How can we bring the request history into a useful draft?
Documented Request Management skills include request summarization and activity response generation.
ServiceNow documentation ↗Check available inputs and permissions. Configure supported fields or integrate missing context; review drafts within the existing approval process.
Illustrative adoption paths, not connected integrations or vendor-generated outputs. Verify edition, version, licensing and configuration. Neither example establishes that this entire workflow is available out of the box. Official documentation checked 10 Oct 2026.