Which project management software does AI recommend?

Jira appeared in 92% of AI-generated answers about project management software — 33 out of 36 responses across 6 models and 6 prompts. Asana appeared in 81%. ClickUp in 75%. Monday.com in 69%. Notion — which is not a project management tool — appeared in 58%.

That Notion number is the first signal that this category is more complex than the SoV table suggests. AI models are answering PM software questions with a mix of dedicated PM tools and workspace tools that have PM-adjacent features. The model’s definition of “project management software” is broader than the vendor category.

The more important finding: Jira was also the first-mentioned tool in 8 of 36 answers — the most of any PM software in this study. ClickUp was close with 7 first-picks, Asana and monday.com each won 6. I ran 6 models: GPT-5.5, GPT-5-nano, Claude Opus 4-8, Claude Haiku 4-5, Gemini 3.1 Pro, and Perplexity Sonar Pro. With these current-generation models, Jira leads on both metrics.

The SoV table in full

Jira: 33/36 (92%). Asana: 29/36 (81%). ClickUp: 27/36 (75%). Monday.com: 25/36 (69%). Notion: 21/36 (58%). Trello: 17/36 (47%). Linear: 14/36 (39%). Wrike: 10/36 (28%). Smartsheet: 6/36 (17%). Microsoft Project: 6/36 (17%). Basecamp: 1/36 (3%).

The top four — Jira, Asana, ClickUp, monday.com — remain the de facto AI shortlist for general PM software queries. If your product isn’t in this group, the model either excludes you entirely or mentions you once in a long list. The gap between the top four and the rest is substantial: from 69% (monday.com) to 58% (Notion), with everything below Notion at 47% or under.

Linear at 36% is interesting: modest overall, but concentrated in the right queries. On the software development team prompt, Linear appeared in multiple flagship model answers where the model was distinguishing between developer-focused tools and general PM tools. Linear’s 36% overall understates its relevance for the dev-team segment specifically.

Why Jira now leads on first-picks as well

With older AI models (Claude Opus 4-5, Gemini 2.5 Flash), ClickUp led first-picks — it had 9 vs Jira’s 7. With the current generation (Claude Opus 4-8, Gemini 3.1 Pro), that reversed: Jira leads 8 to 7.

The driver is the flagship models. Claude Opus 4-8 reaches for Jira first across general, startup, and dev prompts. Gemini 3.1 Pro — which has web retrieval from the broader comparison content ecosystem — also opens with Jira more consistently than its predecessor did. GPT-5.5 and Perplexity were consistent Jira openers in both runs.

ClickUp still leads on Claude Haiku (budget tier) and wins the generic “best PM software for everyone” framing on some models. But the shift in flagship model behavior has erased ClickUp’s first-pick advantage.

The first-pick breakdown: Jira 8, ClickUp 7, monday.com 6, Asana 6, Linear 4, Smartsheet 2, Trello 2, Wrike 1. The two-horse race between Jira and ClickUp is essentially a coin flip at this point — a single model-prompt combination separates them.

The Notion anomaly

Notion appeared in 58% of answers without being a traditional project management tool. This happened because comparison content frames Notion as a PM option for teams that want flexibility over structure. Tech blogs and productivity review sites frequently include Notion in PM roundups — and the model retrieves those articles.

The per-model data: GPT-5-nano (5/6) and GPT-5.5 (4/6) mention Notion frequently. Claude Opus 4-8 (3/6) and Gemini 3.1 Pro (3/6) mention it less. Perplexity sonar-pro (3/6) pulls current web content where Notion appears in many recent PM comparisons.

For dedicated PM vendors, Notion’s 58% is a competitive signal: the AI-defined category boundary is more porous than the vendor category. A buyer asking AI about PM software may receive Notion as a recommendation — and choose it. This isn’t a solvable AI problem; it reflects how comparison content categorizes tools.

Linear and the dev-team niche

Linear appeared in 39% of overall answers, but its distribution matters more than its average. On dev-team queries, Linear appeared alongside Jira frequently — the model positions it as the modern Jira alternative for teams that prioritize speed and simplicity. On general queries, Linear rarely appears.

This is the correct niche positioning for a focused tool. Linear doesn’t need to win generic “best PM software” queries. It needs to win the “project management for software teams” query — and within that context, its AI visibility is meaningfully higher than 39% suggests. The ecommerce platform study showed a similar pattern: niche tools that dominate one prompt type score higher SoV than their overall market position would predict.

Model variation across the PM category

Gemini 3.1 Pro showed the most consistent broad coverage: 6/6 on Jira, Asana, ClickUp, and monday.com. Every query, all four top tools. Its web retrieval pulls from a wide comparison source set and produces comprehensive answers that name every major tool.

GPT-5.5 was similarly consistent: 6/6 on Jira and Asana, 5/6 on ClickUp and monday.com. Claude Opus 4-8 was precise — 6/6 Jira, 4/6 each for Asana and ClickUp. Claude Haiku (budget tier) showed the most variance: 5/6 on Asana but only 3/6 on Jira, the inverse of every other model.

For PM tool vendors: Claude Haiku’s deviation from the consensus is notable because it’s the most cost-efficient model buyers are likely to interact with via chatbot integrations. A tool that underperforms on Haiku but leads on Opus has different marketing implications than vice versa.

The citation sources driving PM answers

The top cited editorial domains (excluding Google’s internal search cache):

• thedigitalprojectmanager.com: 23 citations
• wrike.com: 19 (own domain)
• clickup.com: 16 (own domain)
• en.wikipedia.org: 9
• monday.com: 8 (own domain)
• reddit.com: 8
• siit.io: 8
• zapier.com: 7
• goodday.work: 6
• project-management.com: 6

thedigitalprojectmanager.com at 23 citations is the single most-cited independent editorial source for PM software. If your product isn’t accurately described and positively featured there, you’re missing the most important editorial intermediary in this category.

Wrike’s own-domain citations (19) are notably high given its 28% SoV — its content strategy is retrieving well, but that content isn’t converting into model recommendations at the same rate. ClickUp’s 16 own-domain citations correlate with its strong recommendation position. Reddit at 8 citations reflects Perplexity’s tendency to pull community content — and a signal that PM community discussions on Reddit are shaping AI answers.

What this means for software buyers

The AI shortlist for project management software in 2026 is: Jira (for engineering, and now the general default on flagship models), ClickUp (strong general-purpose, one first-pick behind Jira), Asana (balanced option), monday.com (visual/workflow-focused). Everything else appears as secondary when the query has a specific constraint.

If you’re not on that shortlist, AI recommendations aren’t helping your pipeline. The methodology for measuring your own position and improving it is covered in how to measure AI share of voice.

For comparison: the CRM version of this study showed HubSpot at 97% with a dominant first-pick lead. The accounting software study showed a three-way SoV tie at 86% with QuickBooks winning first-picks 67% of the time. The ecommerce platform study showed Shopify at 97%. The PM category is the most competitive: the top two tools are separated by one first-pick across 36 answers.

The practical next step is tracking AI citations — thedigitalprojectmanager.com at 23 citations is the most actionable target in this category. For the strategic context on how citation targeting fits into a broader AI visibility strategy, see generative engine optimization and how to measure AI share of voice.