# How to Break Into a SaaS Category's AI Shortlist > In September and again in October, four of project management's five places in ChatGPT's and Google's answers went to the same products; in customer relationship management (CRM) software, two. Where a newcomer can still get in, before you fund a way in. By Vlad Shvets, Partner @ Qvery · Published 2026-09-22 · Updated 2026-11-09 · 6 min read Canonical: https://qvery.ai/blog/break-into-saas-category-ai-shortlist/ If your software never appears among the options buyers ask AI to compare, your sales team may never reach those buyers. Plans to get a product named in answers like these (a quarter of comparison pages, a push for reviews, pitches to the sites that publish best-of lists) are often checked against one question typed into ChatGPT and a screenshot. Half of a marketing lead's time ($254,000 a year) and half of a writer's ($127,800) for a quarter costs about $47,700 at the Bureau of Labor Statistics’ mean pay with benefits, before any outside spend. - **4 of 5** project management places taken by the same products in every asking, on both engines, in September and October - **2 of 5** CRM places taken the same way Source: Qvery research: the same buyer questions in project management and CRM, asked on ChatGPT and Google's chat-style answers, September and October 2026 > At Qvery, we help software companies stop spending on visibility plans against seats that are already taken, and keep their marketers on the pages that win them, with custom AI agents that ask your buyers' questions every day, show who gets named instead of you and draft the fixes your team approves. In Gartner's 2025 survey of B2B buyers, 45% had used generative AI during a recent purchase, mostly to gather information on vendors and products. ## What a Quarter Buys Against a Locked Place A shortlist here is the handful of products an assistant names when a buyer asks it to compare options. Over many buyer questions it splits into seats: names that come back every time, and places that different names take in turn. One answer is a draw from that list, never the list. In project management on Google's chat-style answers there was no open seat at all: a newcomer there has to push out one of the leaders. On ChatGPT, in both categories, one seat rotated among a short pool of known rivals, and that place is the realistic target. A month later the places both engines agreed on all survived; one CRM place on Google's chat-style answers changed hands. GEO Agent: https://qvery.ai/geo-agent/ ## Where a Newcomer Can Still Get In When an October answer linked sources for a product it listed, ChatGPT linked that product's own website **98.95%** of the time; Google's chat-style answers did **5.14%** of the time and leaned on lists, comparisons, threads and videos. ## What Qvery’s Agents Watch Every Day The asking is the part worth handing to software: the GEO Agent records the answers, and the agents work from those records and your own list of the names that count as your product, not from third-party pages. For each buyer situation where a rival is named instead of you, the agents draft the page fix, and for each list that leaves you out, the note to its editor, all in one list for your marketing head. What the agents may read and do is set in a written agreement, your data is processed where that agreement says, and every action the agents take is logged. The calculator prices that page and editor work for a firm of about 300 people: two marketing people, each spending about 16% of the week on it. At about $127,800 a year each, the Bureau of Labor Statistics’ mean pay for marketing specialists with benefits, that is about $40,896 a year. With the agents taking about 55% of it, at about $300 a month to run, it costs about $22,003, saving about **$18,893** a year and 366 hours; $10,000 of build pays back in about 6.4 months. Each figure is our assumption, to replace with your own. (An interactive calculator on the page: roi-situation-page-fixes.) > 📌 **What stays with your marketing head** > - Choosing the categories and the buyer questions, every disputed call on which names count as your product, and the decision to fund the plan. > - Opening the pages linked under the answers that matter, and approving every note before it goes out. ## Bring Us the Plan on Your Desk Bring the plan and your category's buyer questions. We have the first agents running within two weeks once access is in place, and your marketing head approves each fix. Have Qvery Build Your Market-Visibility Agents: https://calendly.com/qvery-vlad/30min ## Frequently asked questions ### Does being named in an AI answer mean the product was recommended? No. An answer can name a product to compare it or to warn against it. The agents mark how each answer describes a product, so the list keeps warnings apart from recommendations. ### Why do Google's chat-style answers name different software than ChatGPT? The engines agree on the leaders and part ways after them, and they read different pages: ChatGPT leaned on the products' own sites, Google's chat-style answers on lists, comparisons, threads and videos. ### Can a new SaaS product break into a category where the top five never changes? On an engine where every place is taken by the same names each time, getting in means pushing out a leader, and across two months none of the places both engines agreed on changed hands. Look for the engine, or the narrower buyer question, where a place rotates. ### How often should we recheck the shortlist? Monthly while you are testing an entry plan. Between September and October the places both engines agreed on stayed put, but one CRM place on Google's chat-style answers changed hands. The GEO Agent records the answers daily, so each month's read is on the same footing. ### Which pages do the assistants link to when they name software? In our October answers that linked sources for a listed product, ChatGPT linked that product's own website 98.95% of the time; Google's chat-style answers did 5.14% of the time and leaned on lists, comparisons, threads and videos. ### Do the agents open every page linked under an answer? No. The agents work from the addresses in your answer records, so they need no third-party pages. A person opens the few pages that matter to a decision. ### What do the agents draft? The page fixes for the buyer situations where a rival is named instead of you, and the notes to the editors of the lists that leave you out. Your marketing head approves every draft before anything goes out. ### Who decides whether to fund an entry plan? Your marketing head and whoever signs the budget. The agents show which places are open and which are taken; the decision to fund the plan stays with people. ### Is one ChatGPT answer enough to judge a category? No. The names change from one asking to the next, so one answer is a draw, not a shortlist. Judge the category on many buyer questions asked repeatedly on both engines. ### Do we need a list of the names that count as our product? Yes. Answers name the same company several ways, and how those names are merged decides which places look open. Settle the list before anyone counts, and keep it with one owner. ### How many buyer questions does a fair read of a category need? A few dozen, drawn from sales calls and asked several times on each engine, then again the next month. A short list measures the questions you chose as much as the category; the agents ask the same list each month, unless you change it, so the months compare. ### Does a place taken on ChatGPT mean the same on Google's chat-style answers? No. Read each engine on its own: a product can keep its place on one engine and rotate on the other, and the pages each engine links behind the names differ too. ### What should the agents hand over before I approve an entry plan? A page per engine showing the places taken by the same products every time, the places that rotate and the rivals in them, the close calls, and the page fixes and editor notes drafted for the place the plan expects to win. ### What does Qvery build for a software company entering a category? Agents built around your category's buyer questions. They read the GEO Agent's daily records, mark the places that go to the same products every time and the ones that rotate, and draft the page fixes and editor notes for the places a newcomer can take. Your marketing head approves every draft.