• Home  
  • Add a Jira Service Management AI Agent for Tier-1 Support Deflection
- Workflow & Ticket Management Systems

Add a Jira Service Management AI Agent for Tier-1 Support Deflection

Cut Tier‑1 tickets dramatically with an AI agent that actually resolves issues — or keeps the conversation intact when it can’t. Learn how.

tier one support deflection bot

To add a Jira Service Management AI agent for Tier-1 support deflection, configure the Virtual Agent with intents and a Confluence knowledge base. It then handles password resets, access requests, how-to questions, triage and status checks, and hands unresolved issues to human agents with full context. It requires JSM Premium or Enterprise.

What Tier-1 Support Problems Does the JSM Virtual Agent Actually Solve?

Jira Service Management’s Virtual Agent targets the repetitive, low-complexity requests that consume the largest share of Tier-1 support capacity. It handles five core problem areas:

  • Password and MFA resets – resolves credential recovery before it reaches a human agent
  • Access and permissions requests – collects details conversationally and escalates only when approvals are required
  • Policy and knowledge-base lookups – answers process and how-to questions directly from connected sources like Confluence
  • Ticket triage – routes misdirected requests to the correct queue with full context attached
  • Status updates – deflects progress-check messages that otherwise consume agent time

When an issue exceeds what the virtual agent can resolve, it escalates to a human agent with the full conversation history attached.

There is a limit to keep in mind. Surfacing a knowledge base article still requires the employee to read it, judge its applicability, and troubleshoot independently. Many employees give up and submit a ticket anyway, which is the gap between deflection and true resolution. Adopting ITSM best practices like incident management helps ensure the virtual agent aligns with broader service management goals.

Where the JSM Virtual Agent Runs and How Employees Use It

Because employees work across different tools, the JSM Virtual Agent meets them where they already are rather than forcing a single intake path. It runs across multiple channels:

  • Slack – Employees submit requests without leaving work conversations.
  • Microsoft Teams – Useful for organizations running Microsoft 365 workflows.
  • JSM Customer Portal – Employees chat directly within the service experience.
  • Email and web widget – Currently in beta availability.

Regardless of channel, employees type requests in natural language. The agent interprets context, surfaces answers from connected Confluence content, and escalates unresolved issues to human agents with the full conversation attached. Once active, it can also be added to other channels beyond the defaults to extend coverage.

The virtual agent also supports all major languages, dynamically translating content from existing knowledge articles so employees receive consistent assistance in the language they use. It is designed with scalability needs in mind to handle increasing request volumes without degrading performance.

How to Set Up the JSM Virtual Agent With Intents and a Knowledge Base

Setup begins with two foundational components: intents and a knowledge base. Intents handle structured, repetitive requests, while the knowledge base supports article-driven self-service through Confluence. Effective B2B integration improves data accuracy and streamlines partner communications, which helps the virtual agent operate across systems with real-time visibility.

Configure intents

  1. Navigate to Jira Service Management project settings under the virtual agent area
  2. Create an intent using a template or build one manually
  3. Add a display name, description, and confirmation question
  4. Include at least three training phrases to improve recognition

Set up the knowledge base

  • Connect an existing Confluence knowledge base
  • Enable AI answers in virtual agent settings

Together, these components reduce human escalations considerably. When a customer sends a message, the virtual service agent first attempts to match existing intents before falling back to AI answers. Rovo can also detect knowledge base gaps and suggest new article topics based on recent ticket trends.

How Much Ticket Deflection Can the JSM Virtual Agent Deliver?

Results for AI ticket deflection depend on a few factors:

  • Intent coverage – how many request types the agent handles
  • Knowledge quality – whether Confluence articles actually answer common questions
  • Channel availability – deploying across Slack, Teams, and the customer portal increases reach

The native virtual service agent requires JSM Premium or Enterprise to activate, so teams on Standard must upgrade before any deflection measurement can begin. Admins retain full control, as intents and AI answers can be deactivated at any time through virtual agent settings.

Well-implemented ITSM practices can also deliver over 70% cost savings through process optimization and automation.

How the JSM Virtual Agent Hands Off to Human Support Without Losing Context

One of the strongest arguments for adopting the virtual agent is what happens when it cannot solve a problem. Rather than dropping the conversation, it escalates with full context intact. The handoff includes:

  • The complete conversation transcript
  • Pre-filled ticket fields based on collected details
  • A summary of what was already attempted

The human agent picks up exactly where the AI stopped, with no re-entry and no repeated questions. JSM automatically creates the ticket and routes it to the service desk queue, which keeps resolution time lower.

The virtual agent operates 24/7 across support channels, so customers get instant help outside business hours without round-the-clock staffing. This also supports service request management by streamlining workflows. Atlassian has reported a 50% increase in resolution rate through this automation-driven approach.

Steps to Add a JSM AI Agent for Tier-1 Deflection

  1. Confirm your plan is JSM Premium or Enterprise; Standard teams must upgrade first.
  2. Connect your Confluence knowledge base and enable AI answers in virtual agent settings.
  3. Create intents for your most repetitive requests, such as password resets and access requests, with at least three training phrases each.
  4. Deploy across the channels your employees already use, such as Slack, Teams and the customer portal.
  5. Review knowledge base gaps that Rovo detects and fill them with new articles.

Common Questions About JSM Virtual Agent Tier-1 Deflection

Can I turn the virtual agent off if it underperforms?

Yes. Admins keep full control, and intents and AI answers can be deactivated at any time through virtual agent settings.

Does the virtual agent work outside business hours?

Yes. It operates 24/7 across support channels, so customers receive instant help without human staffing around the clock.

What happens when the agent cannot solve a request?

It escalates to a human agent with the full transcript, pre-filled ticket fields and a summary of what was attempted. JSM creates the ticket and routes it to the service desk queue.

Getting Started With Jira Service Management Tier-1 Deflection

Start small: pick a handful of high-volume request types, build intents for them, and make sure the matching Confluence articles actually answer the question. Measure deflection against true resolution, since an article an employee abandons still becomes a ticket.

Disclaimer

The content on this website is provided for general informational purposes only. While we strive to ensure the accuracy and timeliness of the information published, we make no guarantees regarding completeness, reliability, or suitability for any particular purpose. Nothing on this website should be interpreted as professional, financial, legal, or technical advice.

Some of the articles on this website are partially or fully generated with the assistance of artificial intelligence tools, and our authors regularly use AI technologies during their research and content creation process. AI-generated content is reviewed and edited for clarity and relevance before publication.

This website may include links to external websites or third-party services. We are not responsible for the content, accuracy, or policies of any external sites linked from this platform.

By using this website, you agree that we are not liable for any losses, damages, or consequences arising from your reliance on the content provided here. If you require personalized guidance, please consult a qualified professional.