/
Atlassian Jira Rovo AI - Usage

Atlassian Jira Rovo AI - Usage

How we use Rovo to help answer IT Help Desk tickets

Overview

The IT Help Desk uses Atlassian Rovo, an AI assistant built into Jira and Confluence, to help us respond to some service desk tickets more quickly and consistently.

Rovo does not replace human support, but it can:

  • Suggest draft replies to common questions

  • Send an immediate response to try and help you outside normal working hours

  • Find and summarise relevant IT Services knowledge base articles

  • Help us keep answers consistent with our documented processes and policies

Every response sent to you is ultimately reviewed and owned by a human IT Services staff member.

Where Rovo is used

We currently use Rovo in the following areas:

  • ITHD Jira Service Management project (IT Help Desk)

    • Rovo is available to IT Services agents while they work on tickets.

    • Rovo can draft internal notes or public replies based on the ticket content and our documentation.

  • ITHD Confluence space (IT Services Help Desk knowledge base)

    • Rovo uses selected pages in this space as a knowledge source when drafting answers.

    • This includes our customer-facing KB articles (for example those listed on “All articles” and “All Rovo Articles – ITHD Support Resources”:
      All articles
      All Rovo Articles - ITHD Support Resources

  • Integrated channels (where configured)

What information Rovo can use

Rovo is configured to use the following types of information when helping with ITHD tickets:

Rovo always respects existing permissions. It only uses and surfaces information that the staff member prompting it is already allowed to see.

How Rovo answers are generated for tickets

When an IT Services agent is working on an ITHD ticket, Rovo may be used in one of two ways:

1. On-demand assistance (manual use by agents)

An agent can ask Rovo questions such as:

  • “Summarise this ticket and link to any relevant ITHD KB articles.”

  • “Draft a reply explaining how to connect to the Cambridge VPN and link to our KB.”

Rovo will then:

  1. Read the ticket details (summary, description and comments).

  2. Search the configured knowledge sources (ITHD/ITSDOC spaces and other allowed sources).

  3. Propose a draft answer or internal note.

The agent then:

  • Reviews and, if needed, edits the draft.

  • Ensures it is appropriate for the specific user and situation.

  • Sends the reply from the ticket, taking responsibility for the final content.

2. Automated suggestions (via Jira Service Management automations)

Where configured in the ITHD project’s automation rules, Rovo can be triggered automatically. For example (illustrative pattern based on Atlassian’s guidance):

  • Trigger: “When a new ticket is created in ITHD”

  • Action: “Use Rovo agent” to analyse the request and suggest an initial diagnostic or KB link.

  • Follow-up action: Add a comment or internal note using Rovo’s suggested text.

Automated agents behave as described in Atlassian’s documentation:
https://support.atlassian.com/rovo/docs/agents-in-automations/

In our environment:

  • The rule runs under the permissions of the rule owner (an IT Services admin).

  • Any comments generated are clearly attributable (for example, to an agent identity, with a note that an AI agent drafted the content where supported).

  • Tickets are still monitored and updated by human agents, who can override or correct any AI-generated content.

  • AI ticket responses are designed to be sent outside normal working hours, so that you may receive immediate help and advice before one of our team are available to respond.

What Rovo will and won’t do

Rovo will:

  • Suggest replies that:

    • Are based on our own documentation and known processes.

    • Try to use clear, concise language suitable for staff and students.

  • Highlight and link to relevant KB articles so you can read more detail.

  • Help agents keep responses consistent across similar tickets.

  • Summarise complex or long-running tickets for quicker handovers within the IT team.

Rovo will not:

  • Directly “act” on your devices or accounts (it can’t log into systems, change passwords, or modify settings itself).

  • Make its own decisions about approval, access, or security exceptions.

    • These decisions remain with human staff and follow existing policies.

  • Override University or Department policy. If Rovo’s draft conflicts with policy, staff will correct it before sending.

  • See information outside the permissions of the staff member using it.

Data protection, security, and accuracy

  • Permissions and access

    • Rovo operates within the Atlassian Cloud environment managed by IT Services Physics.

    • It honours all existing Jira and Confluence permissions: if a person cannot view a page or ticket, Rovo will not use that content when responding on their behalf.

  • Use of generated content

    • Responses suggested by Rovo are treated the same way as any other draft text:

      • IT staff review and may edit them before they are sent.

      • Staff remain responsible for ensuring that the final message is correct and appropriate.

    • We may reuse or refine good Rovo-generated explanations as formal KB articles where appropriate.

  • Limitations and errors

    • As with any AI system, Rovo can occasionally:

      • Omit relevant details.

      • Misinterpret a request.

      • Propose wording that is technically correct but not well suited to a specific case.

    • If you ever believe a response to your ticket is incorrect or unclear, please reply to the ticket to tell us – a human agent will review and clarify.

  • Atlassian’s AI architecture

    • Atlassian has adopted a mulit-model AI approach using leading LLM providers such as Anthropic, Google, and OpenAI, with specialised use of (Meta) Llama and Mistral models.

    • A “teamwork graph” is also utilised, which represents Atlassian’s knowledge network and context model applied as part of the wider LLM approach.

    • Atlassian has a “zero-day rentention” policy with these providers, and as part of our subscription agreement our data is not used to train these models.

For more information on how Rovo works at a platform level, see Atlassian’s documentation:

How to tell if Rovo has been used to help with your ticket

We aim to be transparent when AI assistance has been involved. Depending on the channel and configuration, you may see:

  • A note in the ticket comment indicating that an IT Services agent used an AI assistant to help draft the response, or

  • Phrasing that clearly references our KB articles and links to them.

In all cases:

  • An IT Services staff member remains responsible for the content of the reply.

  • You can always ask for clarification or more detail.

Feedback

We’re actively monitoring how Rovo affects ticket quality and response times.

If you have comments or concerns about AI-assisted replies, please:

Your feedback will help us tune our Rovo rules, knowledge base, and workflows documented in the ITSDOC space, so that the service keeps improving.