We just wrapped a Service Collection in action: From intake to autonomous resolution session! In this session, three Atlassians shared a specific AI use case in a Service Collection app (think Jira Service Management, Customer Service Management, Assets, and Rovo). The AI skills we presented can help customers find answers, create tickets (when appropriate), and triage and route tickets to a team that can resolve them.
Here’s a recap of the session, plus some extra resources to help you implement these solutions on your team today. Bookmark this page, test the agents and resources in your service space, and join the discussion to share how you’re using AI throughout Service Collection!
Did you attend the session? Let us know what you thought by completing this quick, anonymous survey!
────────────────────────────────────────
Help customers get answers quickly with a Rovo self-service agent
@Silvia Davis, a Senior Technical Product Marketing Manager at Atlassian, demoed a Rovo self-service agent for Jira Service Management that helps customers get answers quickly through a natural conversation and find a smooth path to human help when needed.
Service teams can deflect common questions to the self-service agent, improve the quality and consistency of responses from the Rovo agent, and make sure escalated requests include useful context from the conversation.
The Rovo agent also gives teams a practical way to improve self-service over time by identifying knowledge gaps, creating new KB articles, and tuning the Rovo agent to the request types that matter most.
Tips for starting out
- Begin with a few high-volume, low-risk customer questions. These should have clear knowledge base coverage. Avoid broad or ambiguous use cases.
- Update your knowledge base. Make sure your knowledge base is up-to-date and written for help seekers.
- Include skills in the agent configuration. Use the Jira Service Management Create Request skill for portal scenarios.
- Make request forms simple and user-friendly. Too many required fields, unsupported field types, or confusing options can prevent a smooth handoff.
Try out the Rovo self-service agent
You can configure the self-service agent from the AI agents tab in Jira Service Management space settings. The agent can be published out-of-the-box, or you can make tweaks to its instructions or configurations.
Be sure to test the full customer journey once you’ve published the portal agent! Ask the agent a question and see how it answers, raises the request, and follows up with the service desk agent.
NOTE: You must be a space admin of a Jira Service Management space with a help portal; have Rovo activated on the site and access to Rovo Studio; and, have relevant knowledge sources connected.
👉️ Check out Silvia’s resources when you’re ready to get started!
────────────────────────────────────────
Triage and route tickets faster with a custom agent
@Gaurav Gupta, a Senior Solutions Engineer for Service Management, demoed an automated ticket triage solution combining Jira automation and a custom Rovo agent. Service desk agents just starting their day often have a ton of tickets in the main queue that need to be properly triaged and sorted. Enter the Ticket Categorizer agent. With this agent, service desk agents can reduce initial categorization and routing time. Imagine starting the day with a clear backlog!
Here’s how it works: As soon as an incident or request ticket is created, Jira automation triggers the custom Ticket Categorizer agent. (This agent is built in Studio--more on that later.) It analyzes the raw summary, description, and metadata against an organization-specific categorization taxonomy. Then, it automatically assigns the exact Category, Subcategory, and Item Type, routes the ticket to the right resolver team and assignee, and posts an explainable comment with its reasoning and confidence level.
Tips for starting out
- Use a categorization framework. Make sure the Ticket Categorizer agent returns a structured format. This will help Jira automation reliably parse smart values into field updates. (See Step 1 of the Ticket Categorizer Setup & User Guide and use the Categorization Framework template.)
- Monitor audit logs regularly. Use Jira automation audit logs to test and verify if a flow triggered correctly and parsed the correct smart values to live fields.
- Use humans for review. Route low-confidence or uncategorized tickets to a fallback queue for human review.
- Use predefined categories. Restrict the agent to only choose from predefined categories.
Try out the custom Ticket Categorizer agent
Create an agent from scratch in Rovo Studio. Start out by testing it manually to see how well it routes and categorizes tickets. Once you’re feeling confident, try building the automation so the agent runs automatically.
👉️ Check out Gaurav’s resources when you’re ready to get started!
────────────────────────────────────────
Resolve requests faster with Request Resolver
@Steve Colman, a Senior Solutions Engineer for ITSM, closed the session with a demo of Request Resolver, an out-of-the-box agent for Jira Service Management. Once tickets are categorized and routed to the appropriate team, a service desk agent starts working on resolving the ticket. Usually this happens manually by reading the ticket, finding the right runbooks or knowledge base articles, drafting replies, updating fields, and moving the ticket through the workflow.
Request Resolver generates a resolution plan for selected service request types, using connected knowledge sources such as knowledge base articles and runbooks. It can help tackle those repetitive, low-complexity requests, freeing up more time for service desk agents to focus on more complex tickets.
Tips for starting out
- Choose the right request types. Start with 1-2 that are frequent, low-risk, and are handled similarly every time.
- Start with supervised mode. This lets your team review and validate plans before executing them.
- Add relevant knowledge. Make sure it has plenty of articles, runbooks, and other guides relating to your chosen request types.
- Configure tools for external actions. These allow Request Resolver to take actions outside of Jira.
Try Request Resolver in your service desk
Like the self-service portal agent, Request Resolver can be configured in Jira Service Management space settings. Let your team know which request types you’re enabling and show them how to use Request Resolver on those tickets. Monitor the results over a few weeks. If you’re getting reliable, consistent results, consider switching to autonomous mode or trying even more request types.
👉️ Check out Steve’s resources when you’re ready to get started!
────────────────────────────────────────
🌟 Share a comment below with your own before-and-after experience. Did you try one of these solutions? All three? Are there any adjustments or customizations that have worked well for you? Let us know!
Keep learning
