Linear's CEO Karri Saarinen declared that issue tracking is dead, and shipped Linear Agent to replace it with context-aware agents. Atlassian went the other way, launching Rovo Dev in Jira, which is AI that fits into the sprint ritual instead.
Both are right about the direction: delivering context to agents is the bottleneck for AI-native code delivery. For most enterprises, the practical answer for the near future is boring: put your context in your sprint tickets, in Jira, Linear, or wherever your business already runs.
Why the issue tracker wins as the context layer
Every alternative has a reach problem: Context in Claude sessions is ephemeral and siloed per engineer. Context in people's heads isn't easily accessible and leaves the company when they do. Context in the repo's Markdown files or GitHub Issues is invisible to anyone who doesn't open the repo, which is most of your business. Context in SharePoint or Confluence reaches everyone, but you can't see the state of the work at a glance. They could work, but issue trackers do this better.
Jira and Linear win because everyone already reads them. If your context lives there, it's readable and editable by the whole business, with every change tracked. This naturally includes your agents, which can execute against the context in parallel with your developers. And every change is versioned, so agents see what's already been tried and get to the right solution faster, saving some precious tokens.
Enterprises can't shift overnight but eventually will
Most large organisations aren't there yet. Sprints and issue tracking are the reality, and proper context layers have not been established. Shifting the delivery model takes time, not only because it takes time to change behaviour, but also because stakeholders need some way to plan and track progress.
Small and technical teams are the exception: with context in the repo, they'll outrun anyone who bureaucratises them into tickets. These teams will end up somewhere close to Linear's vision with agents absorbing procedural work, humans on taste and judgment. But most enterprises will run sprints for years yet, and the ones treating the tracker as their context layer, not just their ticket queue, will capture most of the AI upside without waiting for the org to change shape.
Interested in how we help teams put context to work across the SDLC? Get in touch.