By
Austin Coker
Reading time: 15 min
Last updated: November 6, 2026.
Buyers are building shortlists inside AI answers before they ever reach your site. If ChatGPT can’t find you, can’t verify you, or describes you inconsistently, it recommends the competitor whose story is easier to repeat. You can fix that, but it takes a program run across your own site, review platforms and the third-party pages ChatGPT trusts. A single blog post won’t do it.
Below you’ll find how ChatGPT decides which SaaS products to recommend, the prompts that build a shortlist, why your competitor shows up and you don’t, and what a GEO agency engagement actually includes, month by month, with cost and the questions to ask before you sign.
ChatGPT recommends SaaS products it can find, verify and describe consistently across the sources it searches: review platforms, comparison and alternatives content, integration and pricing pages, community threads, and editorial lists. To get recommended by ChatGPT, make sure OAI-SearchBot can crawl your site, publish specific pages for the prompts your buyers actually use, and build third-party mentions that describe you the same way. A GEO agency runs that as a program: prompt tracking, source gap analysis, page builds and earned mentions, measured on pipeline.
More than 1.2 billion people use ChatGPT each week (source: OpenAI, Oct 7, 2026). Your buyers are part of that number.
G2 surveyed 1,076 B2B software buyers in March 2026. 51% now start research with an AI chatbot more often than Google, up from 29% in G2’s 2025 report. AI chatbots are the top source influencing shortlists, ahead of review sites, analyst firms and vendor websites. 85% think more highly of a vendor when AI includes them in an answer. 69% chose a different vendor than they first planned because it was part of the chatbot’s recommendation (source: G2, Apr 15, 2026).
The last number should get your attention. One in three buyers in that survey bought from a vendor they had never heard of before. ChatGPT is introducing your competitors to your prospects.
Gartner found the same pattern from a different angle. In a survey of 646 B2B buyers, 67% prefer a rep-free buying experience and 45% used AI during a recent purchase (source: Gartner, Mar 9, 2026). The research your sales team used to shape now happens before the first call.
When a buyer asks for software recommendations, ChatGPT may search the web on its own. It sometimes partners with other search providers. When it does, it typically rewrites the buyer’s question into one or more targeted queries sent to those providers. It may then send more specific follow-up queries. OpenAI lists Microsoft among the providers whose privacy policies apply (source: OpenAI Help Center).
So a prompt like “best onboarding software for a 50-person team” doesn’t get answered as one search. It gets split into several narrower searches about features, pricing, segments and reviews. Your pages need to answer those narrower questions directly. We break down how this works in our guide to query fan-out.
OpenAI uses a crawler called OAI-SearchBot to surface websites in ChatGPT search. Sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links (source: OpenAI crawler docs). OpenAI also says your host or CDN needs to allow its published searchbot IP addresses.
GPTBot is a different crawler. It is used to train foundation models, and the two settings are independent. You can allow OAI-SearchBot and block GPTBot. After a robots.txt change, OpenAI says it takes about 24 hours to apply to search.
A security rule or a bot-management default on your CDN can block OAI-SearchBot without anyone on the marketing team knowing. Check your robots.txt and your bot settings before you spend a dollar on content.
Your homepage says you’re the best. Every vendor’s homepage says that. ChatGPT weighs what other sources say about you.
G2’s research found that citations from a software review site are buyers’ top trust signal for AI answers. Buyers also treat consistency across ChatGPT, Gemini and Claude as a trust signal (source: G2, Apr 15, 2026). If three assistants describe you the same way, buyers believe it. If they contradict each other, buyers move on.
When ChatGPT lists SaaS products, it usually attaches a short label to each one. “Best for mid-market teams.” “Best for Salesforce-heavy orgs.” “Budget option.”
That label comes from how sources describe you. If your reviews, comparison pages and roundups all say you’re built for fast-growing B2B teams, that becomes your label. If they say nothing specific, ChatGPT either guesses or leaves you out. You can influence the label, but only by making the same description show up in many places you don’t control.
Most ChatGPT SaaS recommendations come from a small set of prompt types. Each one pulls from different sources, so each needs a different fix. For a deeper breakdown of prompt families by funnel stage, see our GEO for SaaS guide.
It is rarely one issue. Usually it is a few of these at once:
A GEO agency for SaaS runs the work above as one ongoing program. Here’s what a real engagement includes.
Generic prompt lists are useless. The prompt set should come from your sales call notes, lost-deal reasons, the competitor names prospects mention and the segments that close best. A typical set runs 50 to 150 prompts, tracked across ChatGPT, Gemini, Claude, Perplexity and Google AI Mode.
The point is to track the prompts that sit closest to revenue. “Best [category] for [your best-fit segment]” matters more than a broad educational question nobody buys from.
For each prompt where a competitor gets recommended, the agency records which URLs ChatGPT cites. That shows you exactly where you’re missing: a G2 category page, a specific roundup, a Reddit thread, a competitor’s comparison page.
This is the part most in-house teams skip, and it is where the plan comes from. You stop guessing what to publish and start filling specific gaps.
Before any content work, the agency checks robots.txt, CDN and bot-management rules, and whether key pages render content on the server or only after JavaScript runs. It also adds structured data where it clarifies product, pricing and review information.
These fixes are fast and unglamorous. They also decide whether everything after them can work.
This means comparison pages, alternatives pages, integration pages, pricing pages and use-case or segment pages. Each one needs checkable facts: real prices or starting prices, named integrations, specific segments, dates.
Vague positioning copy doesn’t get cited. A page that says “starts at $X per seat for teams of 10 to 200, integrates with Salesforce and HubSpot” gives ChatGPT something it can repeat with confidence.
This is the bulk of the work and the slowest part. It includes running a review program, getting into editorial roundups and “best X” lists, keeping partner directory listings accurate, answering in community threads where your buyers ask questions, and digital PR.
Ask how each placement is earned. Some are paid listings, some are editorial. Both can be legitimate, but you should know which is which.
GEO works better when it shares data with your other search channels. Google Ads shows which search terms actually convert into demos. SEO ranks the pages built around those terms. GEO turns the same pages into AI citations. That’s one plan with one set of pages, so you don’t end up paying three teams for three disconnected strategies.
ChatGPT also has paid ads now, and they are labeled and separate from organic answers. We cover how they work in our guide to ChatGPT ads.
A screenshot of your brand in one ChatGPT answer proves very little. Answers change by user, session and day.
Useful reporting tracks share of answer across your prompt set, citation share by source, GA4 referral traffic from chatgpt.com, self-reported attribution from a “How did you hear about us?” field, and demos. If your agency can’t tie the work to demos, you’re paying for activity.
We’ll run your buyers’ prompts and show you which sources ChatGPT cites instead of you.
Results vary by category, competition and starting point. No one can guarantee placement in ChatGPT. OpenAI itself says ChatGPT ranks results using multiple factors and that placement is not guaranteed (source: OpenAI Help Center). Here’s what a realistic timeline looks like.
Long-tail prompts move first because fewer competitors have answered them well. The broad category prompts take longer because every vendor in your space wants them.
Published ranges on our SaaS GEO agency review run from $5,000 per month (95 Projects) up to $20,000 to $50,000 per month at larger agencies. Our own retainer for SaaS GEO runs $5,000 to $20,000 per month.
The price moves with the scope of the work:
A low retainer that only covers prompt tracking is a monitoring tool with a monthly invoice. A higher retainer should include page builds and earned mentions, because that’s what moves recommendations.
GEO is a young service, and plenty of agencies have added it to an existing SEO offer with a new service page. These six questions separate real programs from rebrands.
Red flags: a guaranteed #1 spot in ChatGPT, prompt counts with no link to your business, “AI-only” content produced at volume with no real expertise behind it, and llms.txt pitched as the main deliverable.
In-house can work if you already have an SEO lead, writers who can produce comparison and pricing content with real detail, and someone who can earn placements and run PR. Most SaaS teams at $1M to $50M ARR have one or two of those roles, not all of them.
An agency makes sense when you need speed, an existing network for placements, and tracking across several AI engines without buying and managing the tools yourself. If you go in-house, budget for a tracking tool. Our Peec AI vs Profound comparison covers two common options.
Google still matters for B2B SaaS, and you don’t have to choose. GEO and SEO share the same pages. A strong comparison page ranks in Google and gets cited in ChatGPT. For the full picture, see our guides to generative engine optimization, GEO vs SEO and GEO for SaaS, or our SaaS search marketing and AI search services.
The research backs the approach. In a controlled benchmark, content changes raised visibility in generative engine answers by up to 40%, with results varying by domain (source: Aggarwal et al., KDD 2024). That’s a research result, not a forecast for your company.
Our own client work points the same way. A staffing firm, Constant Hire, reached #1 in ChatGPT and generated more than $70,000 in revenue from ChatGPT in four months. That’s a staffing company, not SaaS, but the mechanics of earning citations are the same. For a SaaS example, a B2B financial data platform with a $10,000 per year price point grew demo sign-ups 40% in five months after shifting its content from blog keywords to commercial terms, with 3.3x more top-3 keywords.
At 95 Projects, we help B2B SaaS companies between $1M and $50M ARR get found and recommended across Google and AI search. You talk to the senior person doing the work, not a junior team. We run SEO, PPC and GEO as one program, and we measure it on revenue and pipeline, not ranking screenshots. If competitors are showing up in ChatGPT and you aren’t, book a call.
Austin Coker is the CEO and founder of 95 Projects. He has worked in SEO since 2020, when he built and sold an ecommerce brand that grew mainly through search. He founded 95 Projects in 2022 to help SaaS, ecommerce, and B2B service companies show up wherever their buyers search, from Google to ChatGPT and AI Overviews.
Have questions about working with us? Book a 30-minute strategy call to discuss your goals and see if we’re a good fit.
No-pressure conversation
Clear next steps (if we’re a fit)
Talk directly to an expert
Make sure OAI-SearchBot can crawl your site, publish specific pages for the prompts buyers use, and build consistent third-party mentions on review sites, roundups and community threads. ChatGPT recommends products it can find and verify across several sources. Your own site alone is rarely enough.
In a typical engagement, first movement on long-tail prompts shows up in months four to six. Broad category prompts usually take six to twelve months. Results depend on your category, competition and starting point, and no one can guarantee placement.
No, you can’t pay for placement in ChatGPT’s organic answers. ChatGPT ads exist, but they are separate and labeled as ads. See our guide to ChatGPT ads for how they work.
A GEO agency builds a prompt set from your pipeline, finds the sources ChatGPT cites for competitors, fixes crawl access, builds the pages AI engines quote and earns third-party mentions. A good one reports on demos and pipeline, not just screenshots.
Published ranges run from $5,000 per month up to $20,000 to $50,000 per month at larger agencies. 95 Projects charges $5,000 to $20,000 per month. Price depends on prompts tracked, content volume, placements and whether SEO and PPC are included.
No. GPTBot is used for training models, and OAI-SearchBot is used for ChatGPT search. The settings are independent, so you can block GPTBot and still allow OAI-SearchBot. Blocking OAI-SearchBot is what keeps you out of ChatGPT search answers.
Track share of answer across a fixed prompt set, citation share by source, GA4 referral traffic from chatgpt.com, and self-reported attribution from a “How did you hear about us?” field. Tie all of it back to demos. One screenshot of one answer doesn’t tell you much.
