nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

AI Risk Assessment is now generally available in Jira Service Management!

Jack Yu

Hello Atlassian Community 👋

We’re excited to share that AI risk assessment is now generally available for Service Collection Premium and Enterprise customers! This new capability brings AI-driven risk scoring directly into your change issues, so approvers can quickly understand how risky a change is before it goes live and take action to prevent incidents from occurring.

Today, many teams still rely on manual judgment, scattered context, and tribal knowledge to assess change risk. With AI risk assessment, Rovo does the heavy lifting for you by analyzing a wide range of technical and operational factors and surfacing a clear, explainable risk profile right where you work.

Read on to learn more about the new capability!

What's launching

With this general availability release, AI risk assessment is available on Jira Service Management change issues (available in Service Collection Premium and Enterprise plans) and delivers:

1. An AI-generated overall risk score for every change

For eligible projects, change issues now include an AI-powered risk score (e.g. low / medium / high) that summarizes the likelihood that a change could lead to incidents or service disruption.

Under the hood, the model evaluates 10+ risk parameters across two core categories:

Technical risk: Includes failed deployments, sub‑optimal implementation patterns, lack of rollback or test plans, recurring incidents tied to similar changes. Current CI/CD tools supported are GitLab, GitHub, and Bitbucket.
Operational risk: Includes scheduling conflicts, freeze/maintenance window clashes, service and asset dependencies, rollback complexity, infrastructure changes that may introduce downtime.

Note: In future releases, we will be adding additional risk parameters for compliance and business risk.

2. Transparent “why” behind the score

One consistent feedback from customers is that AI results build trust only when the evidence is visible.

That’s why every AI risk assessment includes:

A breakdown by category (technical, operational), with each category’s risk level.
Evidence and drivers behind the score (for example, “3 failed deployments in the last 60 days for this service,” “missing rollback plan,” or “open P1 incident on a dependent service”).
Clear labeling so it’s obvious what’s powered by AI and what’s coming from existing configuration and historical data.

This makes it easier for CAB members and approvers to defend decisions in audits, post‑incident reviews, and leadership conversations.

Screenshot 2026-03-23 at 10.11.11 AM.png

3. Recommended mitigation steps to reduce risk

AI risk assessment isn’t just about flagging problems—it also suggests next best actions so teams know what to do next.

Based on the underlying signals, Jira Service Management surfaces mitigation recommendations such as:

Add or strengthen a test plan, rollback plan, or implementation plan
Re‑evaluate change timing due to conflicting changes or freeze/maintenance windows
Investigate and resolve open incidents or PIRs related to the affected services or assets
Review security, privacy, or accessibility gaps that appear to be out of alignment with your documented controls
Consider the business impact (e.g. high‑tier services or changes aligned to revenue‑critical systems)

Screenshot 2026-03-23 at 10.12.34 AM.png

Why it’s important

AI risk assessment helps teams prevent incidents before they happen by automatically surfacing risk signals from existing operational data and guiding approvers with clear, explainable recommendations. This not only reduces manual toil for CABs and change approvers, but also strengthens trust in AI and advances your overall ITSM maturity and operational resilience.

Prevent incidents, don’t just react: AI risk assessment proactively surfaces risk signals from past changes, incidents, and deployments before approval, helping CABs and approvers spot “hidden” risks. This keeps teams focused on the riskiest changes instead of treating every request the same.
Reduce manual toil for CABs and change approvers: By automatically scoring risk using existing Jira and service data, AI removes much of the manual digging through logs and histories. It adds contextual analysis and concrete mitigation suggestions so approvers can move faster without starting from scratch.
Build trust in AI with transparent evidence: Each risk score is backed by visible, auditable drivers and references, presented in clear, human‑readable language aligned with Atlassian’s AI guidelines. The experience makes it explicit that AI is advisory, ensuring humans remain firmly in control of approvals.
Strengthen operations and ITSM maturity: AI‑driven risk assessment improves change controls, cutting down on recurring, change‑related incidents while accelerating AIOps adoption in Jira Service Management. For regulated and high‑stakes environments, explainable scoring and embedded guidance materially boost both resilience and compliance posture.

Designed for explainability and control

With the general availability of AI risk assessment, we’ve standardized on a clear set of risk categories (technical, operational) to give admins and teams a consistent mental model, ensured project admins can control AI risk scoring and choose which assessment parameters to include, and provided a fallback experience that still surfaces existing conflict and incident insights even when AI is disabled. For more details on how we are building AI features responsibly, check out this page.

Share your feedback with us!

We’d love to hear how AI risk assessment is working for your team and where we should go next (e.g. deeper integration with third‑party CI/CD tools, more advanced prevention dashboards, or additional risk categories).

Please use the comments thread below to share:

How AI risk assessment is influencing your approval decisions
Any false positives or blind spots you’re seeing
What additional signals, explanations, or actions would make this even more useful in your environment

We’ll be actively monitoring feedback here and using it to guide future improvements for change management in Jira Service Management!

Comments

Harrison Ponce

Hi @Jack Yu , thanks for the article! This sounds interesting, but we already have a "Risk Summary" tab in changes and am wondering how different or similar this is from that? Does it replace it?

Also, I don't see this AI Risk Assessment in our Change work types. I don't see it as a feature to enable in the Jira space settings either. Am I missing something?

Tapiwa Samkange

Hi @Jack Yu - thanks for sharing. Does the feature also advise on non-code / software related issues? A quick example off the top being a request to send a laptop to a user who is travelling. A potential risk being the user is at a specific location for a limited amount of time and so will likely not received the package in time?

Troy Anderson_htyevj

Please specify up front in the communications which features are premium and which are actually available to my company/those of us not on Premium.

Troy Edmonds

We do risk and residual risk management on 6x6 matrix with no AI now in Jira, so I hope this can improve on what we can do

Marion Lepmets _SoftComply_

This is a fantastic addition to Jira Service Management! Bringing explainable, AI-supported risk assessment directly into change management can help teams identify potential issues earlier and make better-informed approval decisions.

It also creates a valuable connection between individual change risks and the organization’s wider risk landscape. Technical and operational findings can feed into broader security, compliance and operational risk management, for example, by linking them to organizational risks, controls, owners and mitigation actions.

This complements the work we support with SoftComply Risk Manager Plus, helping organizations manage those wider risks systematically in Jira.

Exciting to see risk-based thinking becoming increasingly embedded in everyday workflows across the Atlassian platform!

Andres Santamaria

This can be very useful, however I'm lacking or not seeing some useful configuration options.

Why are the tasks created for mitigation not linked backed to the original Change ticket? They have a link in the description, but are not using Linked Issues.
Would be good to configure where these mitigation tasks should be created, which space and perhaps which type to use.
Not all Change processess look the same, would be good to choose additional fields as part of the change risk assessment and prioritization.
Tested with a very risky example change that had 16 Risk score (Critical) but was still marked as Medium and even having fields like No roll back, During business hours, critical tier product, on-going incidents and previous outages. How can this AI risk assessment be further influenced so that this would actually be counted as a High risk change?

Rasmus Ersmarker

Two things are getting mixed together in this thread, and separating them helps.

This scores change issues, and Atlassian's wording here is "AI risk assessment is now generally available for Service Collection Premium and Enterprise customers" — so Troy Anderson's point stands: below Premium it is not a feature you have.

Troy Edmonds, what you run is a different job. A 6x6 likelihood/impact matrix with residual risk is a project risk register, not change approval, and nothing in this announcement suggests it replaces one. For residual risk and regulated work SoftComply is the honest recommendation; my own app tops out at 5x5 and has no residual-risk concept at all, by design.

Jack Yu, one real question: does the assessment land on a field — queryable in JQL, readable by automation — or is it only rendered in the approval view? The article does not say, and that decides whether the score can drive anything downstream.

Disclosure: I build one of the Forge apps in this space: FastRisk , so read this as interested.

Welcome!

It looks like you're new here. Sign in or register to get started.

Quick Links

Find more posts tagged with

jira-service-desk-cloud

risk

jira-service-management-cloud

jira-service-management

change-management

rovo

service-collection

rovo-ops