I asked two AI engines the questions a software team asks before they pick an issue tracker. Then I looked at who got credited for the answers. Linear, one of the most loved engineering tools of the last five years, is barely in the room, and where it is, someone else is doing the talking.
This is a teardown, not a takedown. The product is excellent. The AI search visibility does not match it, and the reason is fixable. Here is exactly what I found, on the record, so you can run the same checks yourself.
What I checked
On 12 June 2026 I ran two buyer queries through Perplexity and one through Google, with no personalization tricks, the plain questions a buyer types.
- “What is the best issue tracking tool for software teams?”
- “Best Jira alternative for modern software teams”
That is a deliberately small sample. A full audit runs every priority query several times across all five engines, because these systems are non-deterministic and one run is a coin flip. But two clean queries are enough to show the shape of the problem.
Finding 1: on the category question, Linear is a footnote
Ask Perplexity “what is the best issue tracking tool for software teams” and the answer is Jira. Not Jira-among-equals. Jira as “the industry standard,” with a full feature table explaining why.
Linear appears once, in a row labeled “Solo / lightweight,” described as “simple, fast, easy to use.” That is the slot AI search has filed Linear into: a cute tool for one person or a tiny team.
Linear runs issue tracking for OpenAI, Ramp, and Opendoor. Calling it a “solo / lightweight” option is not a small miss. It is the engine actively mischaracterizing the product to every buyer who asks the most important question in the category.
This is a specific case of a general trap: being cited is not the same as being recommended. Linear is in the answer and still losing, which is why getting cited by AI is not the win it looks like.
Google is worse. Its results and AI summary for the same query name Jira, GitHub Issues, Basecamp, Asana, and Freshdesk. Linear is not mentioned at all. On the single highest-intent question in its market, Linear is absent from Google’s AI answer.
Finding 2: even when Linear wins, it is not the source
Now narrow the query to where Linear should dominate: “best Jira alternative for modern software teams.”
Here Linear wins cleanly. Perplexity calls it “the best Jira alternative specifically for modern software teams,” praises the speed, the developer experience, the opinionated workflows. A great answer for Linear.
Look at the citations under it. YouTube. Reddit. A site called taskrhino. Monday.com, a direct competitor. Across both queries, the pages credited for the answer were review sites, forums, video transcripts, and rival vendors.
The number of times linear.app was cited as a source, across both answers: zero.
That is the real finding. Linear is not losing AI search. Linear has handed the microphone to other people. When a YouTuber and a Reddit thread happen to say nice things, Linear wins. When the category query is broader and those third parties default to Jira, Linear gets filed under “lightweight.” The narrative swings on what other people publish, because Linear has published almost nothing that an engine can cite to answer these questions.
Why this is happening
Linear’s content footprint is, by design, product-facing. A beautiful changelog. Deep documentation. Customer stories from impressive logos. A “Method” section on how to build product.
What is missing is the layer AI engines actually pull from for buyer questions. There is no owned page that answers “what is the best issue tracker for a software team at scale.” No honest “Linear vs Jira” comparison written from Linear’s point of view. No data-backed argument that the engine can quote when someone asks the category question.
So the engines do what they always do when the company is silent: they assemble the answer from whoever did write it down. For Linear, that is Reddit, YouTube, review aggregators, and competitors. The product is loved enough to survive that. It is not winning because of its content. It is winning despite the absence of it, and only on the queries where outsiders happen to be kind.
That is a brittle position for a company this good.
What I would fix
This is a clean CITE Method case, because the gap is so specific.
Map the buyer questions. Pull the real questions a team asks before choosing a tracker, ranked by how close they sit to a decision. “Best issue tracker for software teams,” “best Jira alternative,” “Linear vs Jira,” “issue tracking for engineering teams at scale.” That is the target list, and building it properly is its own discipline: how to build a buyer query map for AI search.
Find the gap. For each one, look at who is cited today. The pattern is consistent: third parties and competitors, no owned Linear page. Every one of those is an open lane. This is the same exercise as finding the content gaps competitors are filling, run against your own brand.
Write the pages the engine can cite. Not marketing copy. Answer-shaped, data-backed pages that open with a direct, quotable claim, the way you write content AI engines will cite. An honest Linear-versus-Jira comparison that concedes where Jira wins and is sharp about where Linear wins. A “best issue tracker for software teams at scale” page that puts the OpenAI and Ramp scale story in writing, so the next time someone asks the category question there is a Linear-authored source for the engine to lean on instead of devopsschool.
Use proprietary data. Linear sits on exactly the kind of first-party numbers AI engines cite most: issue throughput, cycle time, how fast teams ship. Original data is the single most citable thing a company can publish. Linear has it and is not publishing it in a citable form.
Then measure it. Re-run the same queries monthly, several times each, across all five engines, and watch the citation rate move. That is the only honest way to measure AI search visibility, and the only way to know the content is working rather than hoping it is.
The honest caveats
Two things, because this is a teardown and not a sales pitch.
This was two queries on two engines on one day. AI answers shift run to run and week to week, so treat it as a snapshot, not a verdict. The real audit is dozens of queries, repeated, across ChatGPT, Perplexity, Google AI Overviews, Copilot, and Claude. The shape would almost certainly hold, but I am telling you exactly what I ran.
And to be clear: none of this is a criticism of the product. Linear is, by most engineering teams’ reckoning, the best in its class. That is what makes the visibility gap worth writing about. When a mediocre product is invisible in AI search, the fix is to be better. When the best product in the category is invisible, the fix is just to write down what is already true, in a form the engines can cite.
That second problem is the one worth having, and the easier one to solve.
If you want to see what this looks like run properly against your own category, with every query and every engine, that is a paid audit, and the method behind it is documented in full on the methodology page.