Linear · AI product · Mid · 45 to 60 minutes

How should Linear speed up issue triage?

Linear is a software product that teams use to plan projects and track issues. Answer this product management case and get it graded line by line against a weighted rubric.

product strategy
AI product judgment
scope definition
metrics
risk management

The situation

You are the PM responsible for issue intake and backlog organization at Linear. Leadership wants a recommendation for the next quarter on whether to invest in an AI powered improvement in this area. As more teams use Linear, new issues pile up across projects, teams, and related work. Someone often has to read each issue, search for similar items, decide where it belongs, and connect it to the right backlog. That work is repetitive and slow, but mistakes create duplicate work, missed ownership, and messy planning. There is growing customer interest in AI inside core workflows, and some competitors are making noisy launches. But Linear is known for speed, clarity, and trust. A bold AI feature could reduce manual triage work, yet bad recommendations or unclear behavior could make backlog hygiene worse. Your job is to decide whether this is the right AI problem to tackle now, what the first version should do, and how much control the product should keep in the user's hands. Assume one cross functional team can work on this for one quarter using existing product data and infrastructure.

What to produce

Write a 1 to 2 page product recommendation memo.

Your answer must

Define the primary user, the exact workflow pain point, and why it is important now.
Recommend whether Linear should prioritize this AI opportunity now, and compare it with at least one simpler or less automated alternative.
Describe an MVP, including the user experience, the level of automation, and what existing product signals or data it would rely on.
Specify success metrics that cover usefulness, trust, and workflow impact, plus how you would validate quality before a broad rollout.
Identify the top risks or failure modes and how you would mitigate them.

Constraints

Assume one cross functional team and one quarter for an initial version.
The first version must fit into existing issue creation and triage workflows.
Use only information already available inside Linear workspaces, such as issue text and existing organizational metadata.
Wrong routing has real planning cost, so trust and reversibility matter.

Format

Use short headings and bullets, and make any assumptions explicit.

How Linear actually played it

Linear describes building an LLM-assisted feature that suggests where incoming issues belong, combining search, ranking and reasoning. The post frames the problem as manual backlog organization that does not scale. Useful as an example of choosing an AI feature by pain point.

Pick AI use cases where manual work does not scale
Combine search and ranking with LLM reasoning
Offer suggestions, not automatic changes, at launch

Built from How we built Triage Intelligence by Yann-Edern Gillet and Matthijs Wolting.