TL;DR
Jira Service Management is now part of Atlassian's Service Collection and is free for up to three agents. Set it up in this order: choose a plan and data residency, design a small set of request types, configure queues and SLAs, build the knowledge base, add automation, connect your channels and the CMDB, then switch on AI features once your knowledge is solid. Most rework comes from skipping the design steps and going straight to configuration.
Jira Service Management is quick to start and easy to get wrong. You can have a portal live in an afternoon, but a service desk people trust takes some design first. This guide walks through the setup in the order we use on client projects.
Step 1: Choose your plan
Atlassian sells Jira Service Management as part of the Service Collection, together with Customer Service Management, Assets and Rovo AI. You can't buy the apps separately in Cloud. The plans differ in ways that matter for IT teams:
- Free: up to three agents. It already includes the portal, email and chat intake, alerts, on-call schedules and basic automation. It's good for a pilot.
- Standard: adds Assets (the CMDB), Rovo agents and search, a branded help centre, audit logs and data residency.
- Premium: adds change management with risk assessment, advanced incident and problem management, the virtual service agent, alert grouping and heartbeat monitoring, plus a 99.9% uptime SLA.
- Enterprise: annual only, for larger organisations needing more automation, Rovo credits and Atlassian Guard.
Pricing is per agent and varies with team size, and Atlassian raised list prices from 13 October 2026, so use its calculator. Only agents (the people resolving requests) need licences; employees raising requests don't.
Step 2: Set up your site and data residency
On paid plans you can pin your data to Atlassian's Sydney region. That covers requests, comments, attachments, Assets, SLA configuration, incident data and virtual agent data. Set this up early, before you load real data. If you're moving from Data Center, which Atlassian no longer sells to new customers and which reaches end of life on 28 March 2029, plan the migration alongside the new design. Our migration guide covers it.
Step 3: Design your request types
This is the step that decides whether the service desk works. Create a service project from the IT service management template, then:
- List the requests people actually make, using your current ticket data.
- Group them into a small number of clear request types, around 15 to 25 to start. Too many and people pick the wrong one; too few and you can't route or report.
- For each, design a short form that captures what the resolver needs, using plain language employees understand.
- Organise request types into portal groups that match how people think, such as "Access", "Hardware" and "Something's broken".
Step 4: Configure queues, workflows and SLAs
Build queues around how your teams work, for example by team, priority or request type. Keep workflows simple: open, in progress, waiting for customer, resolved. Then set service level agreements for time to first response and time to resolution, by priority, using a calendar of your business hours. If you support several states, account for different time zones and public holidays. Our guide to SLAs in ITSM covers this in detail.
Step 5: Build the knowledge base
Connect a Confluence space as your knowledge base and write articles for the top twenty questions before launch. The portal suggests articles as people type, which deflects simple requests. Later, the AI features answer from these same articles, so their quality sets the ceiling for everything that follows.
Step 6: Add automation
Start with the automation that removes the most manual effort:
- Auto-assign and route requests by type and category.
- Send reminders when a request is waiting on the customer, and close it after a set time with no reply.
- Escalate and notify when an SLA is about to breach.
- Create linked sub-tasks for multi-team requests such as onboarding.
Each plan includes a monthly automation allowance, so design rules that run when something changes rather than on frequent schedules.
Step 7: Connect your channels
Let employees ask for help where they work. Jira Service Management takes requests from the portal, email, and chat in either Slack or Microsoft Teams (a site can use one, not both). On Premium and Enterprise, the virtual service agent answers in those channels from your knowledge base, and hands over to a person when needed.
Step 8: Set up Assets
On Standard and above, Assets is your configuration management database. Start with your critical business services, the applications they depend on and the devices behind them, so incidents show their impact. Assets Discovery can scan your network to populate it. Link assets to request types such as "Report a problem with my laptop" so the right device is attached automatically.
Step 9: Set up incidents, alerts and on-call
Alerting and on-call schedules are built into Jira Service Management. Connect your monitoring tools so alerts open incidents, set up on-call rotations and escalation policies, and agree a major-incident process. If you still use Opsgenie, note that its support ends on 5 April 2027; Atlassian provides an in-app tool to migrate it.
Step 10: Switch on AI features
Once the knowledge base and request types are in good shape, turn on Rovo. Its agents can triage and route requests, summarise long threads, draft knowledge articles from resolved work, and help analyse incidents. Rovo usage runs on monthly credits included in paid plans, with extra usage billed from December 2026. Start with a few well-defined uses and measure the results.
Step 11: Train, launch and improve
Train agents on queues, SLAs and the knowledge base. Launch to one department first, fix what you learn, then roll out. After launch, review SLA performance, satisfaction scores and the most common requests every month, and keep adding knowledge articles and automation where volume is highest.
Common mistakes to avoid
- Copying your old tool's categories. Design request types from how people ask, not how the old system was set up.
- Launching without knowledge. An empty knowledge base means no deflection and poor AI answers.
- SLAs without calendars. Measuring overnight and weekends as working time makes every SLA look breached.
- Leaving integrations for later. HR, identity and monitoring connections are where the time savings are.
How Great Minds helps
Great Minds is an Atlassian Silver Solution Partner. We design and implement Jira Service Management, migrate from Opsgenie, Data Center and other tools, and connect it to your HR, identity, monitoring, CRM and ERP systems. For a mid-sized financial services company, our implementation lifted internal satisfaction from 68% to 85% within three months.
Frequently asked questions
Is Jira Service Management free?
Yes, for up to three agents on the Free plan. Larger teams need a paid Service Collection plan, priced per agent.
Do employees who raise requests need a licence?
No. Only agents, the people who work on requests, need licences.
How long does setup take?
A basic portal can be live in days. A well-designed IT service desk with SLAs, knowledge, automation and integrations usually takes a few weeks to a few months, depending on scope.
Can it handle HR and facilities requests too?
Yes. Many organisations run separate service projects for HR, facilities and finance on the same platform, sharing the portal.
Where to start
Export last quarter's tickets from your current tool and group them by what people were asking for. That one list gives you your request types, your automation priorities and your first knowledge articles. If you'd like help, talk to us.