We ran Pixis Visibility GEO audits on leading project management platforms across ChatGPT, Google Gemini, Perplexity, and Claude. The results show that AI recommendations have quietly become a measurable, winnable channel, and that most brands have no idea where they stand on it.
The buying journey moved, and nobody sent a memo
For two decades, a software buyer with a question typed it into Google, skimmed a page of blue links, and clicked through to a few of them. In 2026, a fast-growing share of that journey never even reaches a search results page. Buyers open ChatGPT, Google Gemini, Perplexity, or Claude and ask, in plain language, "What is the best project management software for my team?" or "How does monday.com compare to Asana for a midsize team?"
The model answers with a short, confident shortlist. Three or four names, a reason for each, and no page two. For most buyers, that shortlist quietly becomes the consideration set. There is no scrolling down to result number eight, because in an answer generated by AI there is no number eight. The brands that are named are the ones that are evaluated. Everyone else is simply absent from the conversation.
This is the shift running through every software category right now, and project management is a near-perfect example of it. The category is crowded, the products are genuinely similar on paper, and buyers research heavily before they commit. That combination makes buyers lean on AI to narrow the field, and it makes the question of who AI names an enormous commercial question. So rather than theorize about it, we decided to measure it. We ran the same generative engine audit we use inside Pixis Visibility on several of the leading platforms, to see exactly who AI names, how often it names them, and where each brand disappears from view.
What we found is that the gap between the brands buyers see and the brands they never hear about is large, consistent, and, encouragingly, fixable.
What we mean by AI visibility
Traditional search optimization is about ranking in a list of links. Generative engine optimization, or GEO, is about something different: how often, and in what context, a brand appears inside the answers that AI engines generate. When a buyer asks an AI assistant for a recommendation, the assistant does not return 10 links. It synthesizes an opinion and names a handful of brands. GEO measures whether your brand is one of them.
This matters more for business software than almost anywhere else. When a brand is cited early and often in AI answers, it gains credibility during the evaluation cycle and earns a place on the buyer's shortlist. When it is missing, it never enters the room. The cost of being invisible in an AI answer is far higher than the cost of ranking eighth on a page of search results, because the AI answer is the whole shortlist, not the top of a long scrollable list.
How we measured it
Every brand went through an identical audit: 50 buyer prompts across four AI engines, OpenAI (ChatGPT), Google Gemini, Perplexity, and Claude, for 200 AI responses per brand. The prompts spanned the full buyer journey, from broad awareness questions like "what is a work management platform," to evaluation questions like "best project management software for marketing teams," to high intent comparisons like "monday.com versus Asana." A brand can look healthy at one stage and vanish at another, and where it vanishes tells you where it is losing deals.
From those responses, we produced a single GEO score on a scale of 0 to 100, combining on-site readiness signals such as author expertise, content depth, and structured data with live results from the prompt tests. Alongside it we tracked citation rate (the share of the 200 responses that named the brand), average position when named, and share of voice against the brand's tracked competitors.
Two caveats. This is a June 2026 snapshot, and because AI responses vary between runs, the figures are directional patterns rather than fixed rankings. And each brand was audited within its own competitor set and vertical framing, so share of voice is relative to different peer groups and is not a single head-to-head league table. The value is in the pattern, which held remarkably steady from one brand to the next.
Note: monday.com was audited within its CRM and work management motion, so its tracked competitor set (Salesforce, HubSpot, Pipedrive) differs from the pure project management peer sets used for Asana and Celoxis. It is included here as a directional comparison, not a like-for-like ranking.
Three brands, three very different places in the AI conversation. Asana appears in more than half of all AI answers in its space and leads its competitive set on share of voice. Monday.com shows up in roughly a third of responses. Celoxis, a capable and well-regarded enterprise tool, appears in only one in five cases, and it loses to Smartsheet in 38 instances where a buyer asked a comparison question and Celoxis was simply not named.
Now look at the average position column, because it holds the single most important insight in this whole study. All three brands rank almost identically well when they actually show up, landing around position 2.7 to 2.8 with 77 to 79 percent of their mentions in the top three. The difference between the leader and the laggard is not the quality of the placement. It is the frequency of appearance. The winning brands are not ranked higher when named. They are named far more often. Visibility, not ranking, is the battleground.
The pattern: AI rewards presence, not just quality
Read past the individual brands, and the same three lessons appear in every audit.
The first lesson is that being good is not the same as being visible. These are all credible products, and when AI names them, it ranks them well. Yet the citation rate ranges from 54 percent to 21 percent. AI surfaces the brands whose presence is strongest across the sources it reads and cites, including review aggregators, analyst writeups, comparison content, and cleanly structured pages on the brand's own site. A genuinely good product with a thin or poorly organized footprint gets skipped, and no amount of product quality can rescue a brand the model never mentions.
The second lesson is that the dark funnel is where deals leak. Every brand scored 100 percent on branded bottom-of-funnel prompts, so when a buyer names you, AI represents you well. The leak is in the generic comparison queries, the questions where the buyer has not yet settled on a name. Asana holds 44 percent there, monday.com 34 percent, and Celoxis 28 percent. That is precisely the moment a rival is substituted in, and it is invisible to traditional web analytics because no click or session is ever recorded. Celoxis is losing 38 comparison moments to Smartsheet, all of which happened inside this blind spot.
The third lesson is that early invisibility compounds. Top of funnel visibility, the "what should I even consider" stage, drops off sharply for the lower-scoring brands, from 39 percent for Asana to 15 percent for monday.com to just 7 percent for Celoxis. If AI never names you during discovery, you are not on the mental list by the time the buyer gets specific, and you cannot win a comparison you were never entered into.
Why some brands win, and why the gap is fixable
The most useful finding is what actually drives the difference, because it is neither mysterious nor permanent. Across the audits, the biggest drag on AI visibility was not product quality, nor brand awareness. It was the set of trust and content signals that AI engines can read on a brand's own website.
The clearest example is a cluster of quality signals known as EEAT, which stands for Experience, Expertise, Authoritativeness, and Trustworthiness. Asana scored 43 out of 100 on this measure, and monday.com scored 36 out of 100, both held back by missing author bios, thin external citations, and sparse structured data. In every audit, closing this trust gap was the single largest available lift, and it is a content and markup fix rather than a product rebuild.
Monday.com's technical details illustrate how ordinary these gaps are. Its site is highly crawlable, scoring a perfect 100 on AI crawlability, with all major AI crawlers allowed and a healthy sitemap. Its internal linking is also a perfect 100. Yet its structured data score is 58 out of 100, using only five of the 18 recommended schema types, and its content depth score is just 40 out of 100. The audit found that none of the 50 sampled pages included an author bio, and the average page length was a modest 732 words. None of these is an exotic problem. They are the everyday hygiene of publishing content that AI can parse, trust, and cite, and every one of them is well within a marketing team's control.
That is the real headline of this study. AI visibility behaves like search optimization did twenty years ago. It is a channel you can measure, benchmark against your competitors, and deliberately improve over a period of weeks, not years. The difference is that most brands can quote their Google ranking for a keyword to the decimal point, yet cannot tell you how often ChatGPT names them for their single most important buying question. They are flying blind on the channel that increasingly decides the shortlist.
What brands can actually do about it
The audits did not just diagnose. They pointed to a consistent, practical set of moves, and the highest impact ones are surprisingly cheap.
Strengthen the trust signals first. Add author bios with real credentials to content pages, publish clear about and contact information, and add person and organization structured data. This was the top-ranked quick win in every audit because it directly increases the EEAT score that AI engines use to decide whether a brand is a credible source.
Expand structured data coverage next. Adding schema types such as FAQ, HowTo, Product, and Review markup makes content far easier for AI to extract and repeat. FAQ content, in particular, is associated with materially higher citation rates because engines lift and reuse concise answers almost verbatim.
Then close the funnel gaps you can see. For brands leaking in generic comparison queries, that means publishing comparison and use case content that speaks to the exact questions buyers ask before they name a vendor. For brands weak at the top of the funnel, it means building genuinely educational content that answers the early "what is this and how do I choose" questions, so the brand enters the buyer's mind before the shortlist forms.
Finally, treat each engine as its own surface. Because the four engines diverge so sharply, the brand that measures per engine and fixes the weakest one will pull ahead of competitors who treat AI as a single monolith.
AI visibility is the new SEO, and the clock is running
What is happening in project management is happening in every category, from customer relationship management to cybersecurity to running shoes. Buyers are handing the shortlist to AI, and the models are quietly deciding which brands make it. That makes a brand's presence in AI answers a real, ownable growth channel, in exactly the way organic search rankings were two decades ago.
The brands that treated search optimization seriously and early captured a durable advantage that latecomers spent years trying to claw back. The same window is open now for AI visibility, and it is open precisely because so few brands are measuring it yet. The three platforms in this study are not outliers. They are a snapshot of a category in motion, where a strong product can still be badly underrepresented, and where the fixes are known, affordable, and mostly a matter of deciding to look.
Where Pixis Visibility comes in
Every figure in this article was produced by Pixis Visibility, one platform to track, optimize, and own your brand's presence in the era of AI search. It exists to answer the question that this whole study circles: when AI recommends software in your category, does it recommend you, and if not, why not?
With Pixis Visibility, you can track your citation rate, position, and prominence across more than 6 AI engines, running the same audit behind the numbers above continuously rather than just once. You can benchmark against your competitors and see who owns each engine's default answer, along with every generic comparison where a rival gets named instead of you. And you can find and close the gaps, from the top of the funnel and comparison content that lifts citation rate to the trust and structured data fixes that move the underlying score, going from insight to published content built for AI in hours rather than weeks.
The brands here are a snapshot of one category on one set of runs. Your category is being ranked by AI right now, whether or not anyone on your team is watching. Pixis Visibility is how you find out where you stand, and how you start winning the answer.
See where your brand ranks. Get a Pixis Visibility audit.
Methodology: GEO audits conducted through Pixis Visibility in June 2026. Each brand was tested with 50 buyer prompts across OpenAI, Google Gemini, Perplexity, and Claude, for 200 responses per brand, scored on a 0 to 100 GEO index that combines on-site AI readiness signals with live AI visibility results. A single run was captured per engine, so the figures are point-in-time and directional, since AI engine responses vary between runs. Share of voice figures are normalized within each brand's tracked competitor set and vertical framing and are not directly comparable across brands.