AI agents are about to route around every tool that can't pass 5 structural tests. Here's the diagnostic.
Karri Saarinen, the CEO of Linear, declared issue tracking dead in March. His argument was reasonable. Issue trackers were built for a world where coordination between humans was the bottleneck, and when agents can interpret context directly, that translation step is friction. It was a clean thesis from a designer who had spent his career making the category better.
TL;DR
- AI agents necessitate a reevaluation of traditional software tools, turning issue trackers into strategic infrastructure.
- The bottleneck of human coordination is diminishing as AI agents interpret context directly.
- Tools with established state machines, assignee fields, audit histories, and dependency graphs are crucial for AI agent operation.
- Linear's design, initially criticized, has become a prime example of effective agent substrate due to its clean data structure.
- The market is recognizing the AI value in existing enterprise software, such as Atlassian's Jira, potentially leading to a repricing.
- A diagnostic is needed to determine which tools (CRMs, ERPs, service desks, etc.) will become agent infrastructure and which will be superseded.