Table of Contents
New Jira and Teamwork Graph features help development teams plan, assign, control, and measure the work of humans and AI agents.
Whether I’m talking to customers or Atlassian’s development teams, the message is always the same: the unprecedented introduction of powerful coding agents has transformed software development, but the difficult aspects of it Provision of software are… gasp… still pretty hard.
Teams still need to decide what they want to build and why it should exist. They need to understand what system they are changing and what limitations are important. You need to know what “done” means and whether the issue is safe to send.
This is the reality behind the AI productivity gap. In a longitudinal study we conducted using DX in professional development teams, AI usage increased by 65%, but overall developer velocity did not. The peak was an increase of 15%, with many companies recording an average increase of 10%.
The gap isn’t because models are bad at writing code. That’s because software development was never just about writing code. It’s about translating business goals, strategies and context into working software within a real organization.
For more than two decades, Jira has evolved to serve software teams of every shape and methodology: We are the source of truth about what to build, who is doing it, how it is running, and what is being delivered.
Today, the shape of software teams is changing rapidly.
AI-native software development teams need a new system where the context for agents is a first-class citizen and tasks are delegated to agents while humans control and review. Engineers, product managers, designers, and security teams bring judgment and context to the work.
It’s a new way for people and agents to work together with clear plans, shared context, and confirmation that the outcome is something the team can get behind.
Today we’re announcing new agent product development capabilities in Jira designed for this change. Teams can plan their work with AI, convert intents into agent-ready specifications, assign work to programming agents, monitor sessions, automate technical loops, and measure AI costs relative to output.
Jira started as a bug tracker and now serves as a system of record for the work of millions of teams. And we will continue to evolve to support the AI-native teams of the future.
What AI native software development means
The practical version is as follows: the SDLC must become readable by agents without becoming less accountable to humans.
This means three things.
- The intention must be structured before work begins. An agent needs more than a prompt or a Jira summary. It needs the requirements, relevant architecture, decision history and constraints that the team already knows.
- Choosing the right agent should not lead to divergent processes. A team can use the Cursor IDE for web development, Claude Code for complex backend tasks, a custom agent running in a cloud sandbox for unique codebases, and Jira Coding Agent to cost-effectively automate routine fixes. The workflow should not branch every time the runtime changes.
- Autonomy must remain observable. Agent sessions cannot disappear into terminals, tabs, or separate logs, keeping critical context locked on local devices. Teams need to see what happened, who reviewed it, and what work item started it.
When you do these three things together, the system changes. Agents stop acting like isolated co-pilots and participate in the same SDLC as the rest of the team.
This is where the teamwork graph comes into play. It’s Atlassian’s context layer: a living map of work, code, people, decisions, and dependencies that helps agents understand not only the task, but also the system around them.
Why Jira is the right place for this change
As the majority of coding work shifts to agents, agents require well-defined tasks with rich, explicit context to deliver high-quality code while efficiently managing token costs. And context is almost never in one place, which is why Atlassian developed the Teamwork Graph: to bring together that atomic task in Jira, requirements in Confluence, conversation in Slack, code context from GitHub, and customer insights from Jira Product Discovery.
Jira uses Teamwork Graph context to break big ideas into atomic tasks that agents can work on, and aggregates the context so they can use it at work.
Without context, agents create code that later creates a productivity bottleneck. You solve the ticket too literally. They miss the architectural constraint. You generate a PR that seems plausible until a senior engineer spends an hour unwinding it.
That’s why this launch isn’t just about deploying more agents in more places. It’s about giving agents access to the organizational memory your team already relies on.
The Teamwork Graph provides context. In Jira, this context becomes the workflow: the intent starts there, the agent’s work is assigned there, the session history is recorded there, and the output comes back there for review.
What we are announcing today
Agents’ work gets interrupted in predictable places: vague plans, losing handoffs, and output teams don’t know how to trust. We built for these vulnerabilities.
1. Plan with better context
- Jira planner brings spec-driven development to Jira. For complex projects, Jira Planner draws on your codebase, Jira and Confluence history, and team context to define requirements and generate a structured technical specification in Confluence. The output is human-readable and useful to an agent. One artifact, two target groups.
- Jira for Slack transforms the conversations that create work into context-rich Jira work items. Teams can ask @Jira to create work, capture the nuances of a thread, sync conversation updates as comments, and assign work to programming agents without losing the discussion that shaped the decision. We’re also launching expanded Microsoft Teams features soon.
- Loom video prompts Turn what you show and say into structured instructions that agents can use to complete tasks. Record your screen and discuss the task. Loom captures your screens, clicks, links, and voice prompts and generates an action plan that you can convert into agent-ready Jira work items in just a few clicks.
Delegate the work to the right agent
- Agents in Jira Let teams assign work items directly to Claude Code, Cursor, or GitHub Copilot, with Codex coming soon. The work remains anchored in Jira as a source of truth while context flows to the agent executing the work.
- Jira Coding Agent is integrated into every paid Jira plan. It can take a well-defined work item, leverage business context and code intelligence via the Teamwork Graph, make the change, and return a review-ready pull request without requiring a developer to move to an on-premises environment for routine fixes.
- Agent sessions in Jira: See which AI coding agents are stuck, what’s awaiting review, and what’s completed. Every engineer working in Jira gets visibility into agent sessions running in their areas and repos in a single view, grouped by what needs attention first.
Scale agent engineering with governance
- Coding agent automations in Jira Let teams route routine tasks like bug fixes, vulnerability remediation, test generation, and documentation updates to agents using Jira’s enterprise-class automation rule generator. Engineers are notified when a PR is ready, and every step remains tied to the original request.
- The Agentic Engineering project template helps teams create agent-ready Jira projects in minutes, with pre-configured workflows, status, tracking, and integrations.
- DX AI cost management gives engineers a way to understand the economics of AI development. It unifies spend and token data across tools like Claude, Cursor, GitHub Copilot, and Jira, assigns these investments to teams and projects, and estimates cost per PR in DX.
A system for people and agents
We ran these patterns within Atlassian with our own engineering teams. We found that Atlassian’s Teamwork Graph provides the business context behind many of these features, connecting work, teams, goals, code, and knowledge across the SDLC so agents can act more relevantly and accurately. In internal benchmarking, agents enriched with Teamwork Graph showed 44% more accurate results and used 48% fewer tokens than agents working without this context. Additionally, we have seen a reduction in PR cycle time and less time spent on routine tasks.
“The bottleneck in AI native development is not agent capability, but rather coordination at scale to keep our engineers on track. We’re working with Atlassian to solve this problem: a place where every agent action is visible, governed, and tied to a business outcome.”
Sean Jörg
Agents will change the way software is built. They will not eliminate the need for judgment, context, or accountability.
That’s why Jira needs to evolve. The next chapter of Jira isn’t just about tracking software work. It helps people and agents work together.
Head over to jira.dev today to try it out.
https://www.atlassian.com/blog/company-news/ai-sdlc
