ChatGPT and AI Are Changing How We Research & Buy Software
If you’re a software founder, sales rep, or marketer, you’ve likely had a conversation where someone said they discovered your solution through ChatGPT, Gemini, or another AI model. Instead of a Google search or a review site, they asked an AI assistant a question—and it recommended your company or product.
Naturally, this leads to a follow-up question: How can you get more “leads” from ChatGPT? Is there such a thing as “SEO for LLMs”? And if and how do reviews factor into it?
The Changing Landscape of B2B Software Discovery
AI has casued a seismic shift in the way we discover and evaluate B2B software in 2025.
Just a couple of years ago, buyers relied heavily on Google searches, analyst reports, and review sites like G2 and TrustRadius. Today, AI tools like ChatGPT and Google Gemini are quickly playing an important role in the software buying journey—and this trend will only accelerate.
AI-powered models are now capable of providing vendor comparisons, software recommendations, and even category insights—all without directly visiting review sites.
This raises a big question:
How do AI models like ChatGPT & Gemini determine which B2B SaaS vendors to recommend?
To understand this, we need to explore two key areas:
1️⃣ How AI models rank and recommend products in general
2️⃣ Whether review sites like G2, TrustRadius, and Gartner Peer Insights play a role
Let’s start with a look at AI-driven recommendations in a broader B2C context before narrowing in on B2B software buying behavior.
The Changing Landscape of B2B Software Discovery
B2B software buying behavior has shifted dramatically in 2025. According to the latest Wynter CMO study of 100 B2B SaaS CMOs and VPs (from $50M to $1B+ companies):
- 24% of B2B buyers now start their vendor search with AI tools like ChatGPT and Gemini—a figure that was virtually zero just a short time ago. “ChatGPT and Perplexity are now common starting points for vendor discovery. Leaders use AI to get quick market overviews and initial vendor comparisons.”
- 54% still rely on review sites like G2, TrustRadius, and Gartner Peer Insights to validate what they’ve heard from peers. “They start with peer recommendations, and then 54% use review sites (G2 dominates) and 35% Google.”
- Google is no longer the automatic first stop—just 51% of buyers use it for vendor discovery, and even then, it’s often for validation rather than initial research.
With AI-assisted vendor research growing fast, a crucial question emerges: Do AI models like ChatGPT and Gemini pull data from review sites? If so, how much influence do review scores and rankings have on AI-driven recommendations?
Before we dive into B2B software, let’s look at how AI makes recommendations in a broader context.
Neil Patel’s Study: What Makes ChatGPT Recommend a Company or Product?
To better understand how ChatGPT determines recommendations, SEO expert Neil Patel and his team conducted a 100-question experiment.
The Study Setup
They asked ChatGPT over 100 questions about consumer products, including:
- “What’s the best laptop for college students?”
- “What’s a great credit card when you have a low credit score?”
- “What’s the best electric SUV?”
- “What headphone brand should I buy if I’m watching movies?”
Then, they gave the responses to a statistician and data analyst, who examined 82 different factors to determine what influenced ChatGPT’s recommendations.
Key Findings: The 6 Factors That Matter Most
Their analysis found that of 82 factors, only six key factors strongly correlated with whether ChatGPT recommended a company or product:
1️⃣ Relevance (0.91) – How well the company or product matches the query.
2️⃣ Brand Mentions (0.87) – How often the company is mentioned across the web.
3️⃣ Reviews (0.61) – Customer feedback and sentiment where available.
4️⃣ Authority (0.52) – The perceived credibility of the company.
5️⃣ Age (0.46) – How long the brand has been around.
6️⃣ Recommendations (0.28) – How frequently the product is recommended across sources.
What does ChatGPT think about findings on key factors in how it recommends a product or company?
I was curious to see how ChatPT would respond, so uploaded the screenshot the six key factors and their weightings, and prompted:
“are those factors and their weight accurate? e.g. reviews (0.61%). base it on your algorithm and how you’re trained.“
ChatGPT’s response*
*Note: I work at a B2B SaaS vendor called, Wisedocs. It’s mentioned by name a few times in the ChatGPT output. Consider Wisedocs a placeholder name, representing any business software company.
The factors listed in the image (brand mentions, reviews, relevancy, authority, etc.) are directionally correct, but the exact weightings (e.g., reviews = 0.61%) are not verifiable or necessarily accurate based on how my algorithm works. Here’s why:
How ChatGPT Actually Decides What to Mention
Unlike search engines that rely on direct ranking factors like backlinks, AI models work based on probabilistic reasoning and context relevance. Here’s what truly matters when I generate responses:
1. Relevancy (Most Important Factor)
- ChatGPT’s primary goal is to provide contextually relevant answers based on the user’s query.
- If a product or company aligns with the topic (e.g., Wisedocs for medical record reviews), it is more likely to be mentioned.
- Weight: High (but varies by query).
2. Training Data Prevalence (Brand Mentions & Authority)
- If a brand, product, or company appears frequently in reputable sources, it is more likely to be included in responses.
- This includes news articles, research papers, well-cited blogs, and Wikipedia.
- Example: If users ask about AI-driven document review, I may reference companies with strong industry mentions.
- Weight: High but depends on the quality of sources (not raw volume).
3. Web Search Results (Real-Time Information)
- If my response is based on a live web search, the sources I find influence which companies or products I mention.
- This is dynamic—if Wisedocs is well-represented in recent articles, it will be included.
- Weight: Depends on whether a web search is used.
4. Authority & Credibility
- Sources that are government-backed, peer-reviewed, or from industry experts are considered more reliable.
- Example: If a Forrester or Gartner report ranks claims processing platforms, I am more likely to reference the companies listed there.
- Weight: Medium to high.
5. User Behavior & Prior Queries
- If many users frequently ask about or mention a product/service, it may appear more often in AI-generated responses.
- However, this is not as direct as search engine rankings and does not function like an SEO-driven algorithm.
- Weight: Medium (but indirect).
6. Reviews & Recommendations
- Reviews are considered only if they appear in reputable sources that I can access (e.g., professional review websites, industry reports).
- Unlike Google, I don’t crawl user-generated reviews on sites like Amazon, Yelp, or Trustpilot.
- If a product is highly rated in major industry reports or media (e.g., TechCrunch, Forbes, Gartner), it may gain more recognition.
- Weight: Lower than relevancy and authority, unless reviews are highly cited in authoritative sources.
Are the Percentages in the Image Accurate?
🔴 No – These weights (e.g., “Reviews = 0.61%”) seem arbitrary and not based on OpenAI’s actual methodology.
🟢 Directionally Correct – The types of factors listed (relevance, authority, mentions, recommendations) do matter, but their exact contribution is not publicly disclosed or fixed.
Reality:
- Relevancy and authority matter more than “reviews.”
- Brand mentions matter, but only from credible sources.
- Web search can override prior knowledge in real-time.
- No fixed weightings exist in the way an SEO ranking algorithm functions.
SEO Expert Weigh in: Optimizing for AI
Winning with Content in the Age of AI: So, Uh… How Do We Actually Do This?
Alex Birkett’s take on optimizing for AI, as discussed in a recent 2025 webinar with Wynter (watch here), shares some common ground with Neil Patel’s findings—relevance, brand mentions, and authority seem to play a role in whether AI surfaces a company.
But while Patel’s study puts numbers behind these factors, Birkett approaches it with a bit more caution.
AI search isn’t as direct as Google rankings, and the exact mechanics are still murky.
That said, the “Surround Sound SEO” approach—getting your brand mentioned across multiple reputable sources—appears to be a smart move.
It’s not a perfect formula, but making sure your brand is visible, credible, and consistently referenced in high-authority spaces is one of the best bets we have for now.
AI optimization might not be an exact science yet, but strong content and strategic visibility are still safe plays.
I asked a B2B SaaS CMO and SEO expert (who also teaches SEO at a top university) what they were seeing with AI-driven search, and their response was more of a cautious observation than a definitive answer: “We’re seeing an increase in ChatGPT referrals as well, and it definitely seems like it’s just coming off the back of us ranking a whole bunch and having a solid DA relative to others in the space.”
In other words, even the experts don’t have concrete proof of how AI-driven search actually works—just signals that suggest it’s still tied to traditional SEO fundamentals like ranking well on Google, building strong domain authority, and showing up in enough credible places. Right now, the best anyone can do is connect the dots and adapt, but there’s no hard-and-fast formula for “optimizing for AI” just yet.
What This Means for B2B
Patel’s study focused on B2C products like laptops and credit cards. But what about B2B software? Does ChatGPT rely on G2, TrustRadius, Gartner Peer Insights, or other review platforms when recommending SaaS solutions?
This is where we shift focus to how LLMs treat software reviews and vendor recommendations in B2B SaaS.
What ChatGPT Says About Review Site Data
ChatGPT has confirmed that it does not directly pull data from software review platforms like G2, TrustRadius, or Gartner Peer Insights. However, these platforms can influence responses indirectly through external citations.
Here’s how ChatGPT responded when asked about it directly:
I wanted to revisit a key point from ChatGPT’s previous response about how it decides which companies or products to mention. It referenced B2C review sites, but noticeably did not mention B2B SaaS review sites like G2 or TrustRadius.
So, I followed up with this question:
Prompt: You said, “Unlike Google, I don’t crawl user-generated reviews on sites like Amazon, Yelp, or Trustpilot.” Do you crawl Gartner Peer Inisghts, G2, TrustRadius?
ChatGPT said:
ChatGPT
No, I do not crawl or directly pull data from Gartner Peer Insights, G2, TrustRadius, or other review platforms. However, there are indirect ways I might reference information from these sources:
1. If the Information Is Publicly Available and Cited in Other Sources
- If news articles, blog posts, or industry reports summarize G2, Gartner Peer Insights, or TrustRadius ratings, I may incorporate that information when responding.
- Example: If Forbes or TechCrunch writes an article saying, “Wisedocs is rated as a top AI claims processing platform on G2,” then I might reference that statement.
2. Web Searches May Surface These Reviews
- If I perform a real-time web search, I may return information from pages that summarize G2 or Gartner Peer Insights data.
- However, I cannot access gated or login-restricted content (e.g., inside Gartner’s paid research reports).
3. AI Training Data Does Not Include Crawled User Reviews
- Unlike Google, which indexes review pages, I do not crawl or store dynamic user reviews from sites like G2, TrustRadius, or Gartner Peer Insights.
- This means I cannot “rank” companies based on star ratings or direct user feedback from these platforms.
How This Affects Software Vendors like Wisedocs
- Gartner, G2, TrustRadius Reviews Still Matter
- Even though I don’t pull from them directly, if Wisedocs earns high ratings and strong customer reviews, it may get mentioned in media and industry reports that I do reference.
- High placements in Gartner Magic Quadrants or Forrester Waves will increase credibility in AI-generated responses.
- Press & Blog Coverage of Wisedocs’ Reviews Helps Visibility
- Encourage tech bloggers, insurtech analysts, and media outlets to mention Wisedocs’ ratings from G2, TrustRadius, or Gartner.
- Example: “Wisedocs is a top-rated AI platform for claims automation, according to G2’s latest report.”
- AI tools will recognize this information if it appears in well-cited sources.
- Optimizing for AI and Web Search Together
- Having a strong presence in industry reports + media + review platforms will maximize Wisedocs’ exposure in AI-generated recommendations.
Key takeaways:
✅ No direct crawling of review sites – ChatGPT does not access G2, TrustRadius, or Gartner databases or store user-generated reviews.
✅ Review site data only matters if it’s cited elsewhere – If a publication, industry blog, or analyst report references rankings from G2 or TrustRadius, ChatGPT may surface that information.
✅ Real-time search can surface summaries, but not direct rankings – When web search is enabled, ChatGPT may retrieve articles discussing review sites but does not actively index them like a search engine.
But can we know for certain if it’s true? ChatGPT is known to “hallucinate” and make things up. My take is that’s probably accurate. Even if it is, does that mean other LLMs aren’t trained on it? I asked Google’s Gemini to see how it answered.
What Google Gemini Says About Its Training Data
Google Gemini takes a more vaguely stance than ChatGPT about the use of software reviews sites in its training.
When I asked Gemini directly, it responded:
It does not disclose exact sources of training data due to confidentiality.
- Gemini is trained on publicly available information, but that doesn’t mean it has direct access to review site databases.
- It may have learned about review sites from mentions in articles, blog posts, or vendor websites.
- Like ChatGPT, it does not provide real-time rankings or direct summaries from review sites.
What This Means for B2B SaaS Vendors & Software Review Sites
1. AI Won’t Automatically Recommend Vendors Based on G2 or TrustRadius Scores
- Simply having a high rating on G2 or TrustRadius doesn’t mean AI will recommend your software in responses.
- AI models rely more on brand awareness, industry mentions, and media citations than direct review scores.
📌 Action Item: If you have strong review site rankings, get them mentioned in external sources like analyst reports, media coverage, and industry blogs.
2. AI Prioritizes Trusted Industry Sources (Analyst Reports, Media, and Blogs)
- AI models weigh Gartner, Forrester, IDC, and major tech media (e.g., TechCrunch, Forbes, Digital Insurance) more heavily than review sites.
📌 Action Item:
- Get covered in analyst reports and top media publications.
- Secure expert mentions in industry thought leadership discussions.
3. Review Sites Still Matter—But Less For LLMs Than For Google Search
- 54% of buyers still use G2 and TrustRadius to validate their vendor choices
- Many check review sites after hearing about a vendor from peers or media, rather than starting their search there.
- But they don’t directly influence how LLMs like ChatGPT and Gemini recommend products and services
📌 Action Item:
- Make sure your G2 and TrustRadius profiles are optimized, with recent positive reviews.
- Encourage customers to cite their reviews in LinkedIn posts, case studies, or media features.
4. AI Uses Real-Time Search—But Doesn’t Index Review Site Databases
- If ChatGPT or Gemini runs a web search, it may find articles that summarize G2 or TrustRadius rankings.
- But it does not pull data directly from those sites or rank vendors based on internal scores.
📌 Action Item:
- Write blog posts and press releases discussing your category leadership in review sites.
Final Takeaway: AI Discovery Is Growing—Is Your Brand AI-Ready?
With 24% of B2B buyers now using AI tools like ChatGPT and Gemini to start their search, SaaS vendors need to optimize for AI visibility—just like they once did for Google.
🔹 AI does not rank vendors based on software review site scores.
🔹 Your best reviews only matter if they are cited in external, authoritative sources (and as far as I can tell from my research, that does not include software user review sites, but does include industry analyst sites.) In my opinion, that might make Gartner Peer Insights even more important give its user reviews and publications are on the authoritative gartner.com domain.
🔹 The new buyer journey blends AI, review sites, Google, and peer recommendations.
Want to improve your AI discoverability?
Focus on getting your brand mentioned in trusted industry sources, analyst reports, and media—because that’s what AI actually sees. 🚀
TL;DR: How to Get AI Models to Recognize Your Brand
✅ Get featured in analyst reports and industry media
✅ Turn your review achievement on G2 & TrustRadius into press releases on leading newswires
✅ Continue to create and distribute content that raises your relevancy and authority
AI-assisted buyer research is here. Is your brand showing up when buyers ask ChatGPT for the best software in your category?
Can you optimize a vendor on ChatGPT like you can with search engine optimization (SEO)?
ChatGPT’s Response:
“The concept of ‘optimizing a vendor’ on ChatGPT doesn’t work the same way as SEO does for search engines, but there are ways to improve visibility and engagement if you’re a vendor looking to be recognized or referenced more frequently in AI-driven conversations.”
Why It’s Different from SEO
🚫 No Indexed Web Content – ChatGPT doesn’t browse the web in real time unless explicitly using a web tool. Unlike Google, which ranks content based on backlinks, keywords, and authority, ChatGPT generates responses from its trained knowledge base and external web searches when needed.
🚫 AI Doesn’t Rank Vendors – There’s no ranking algorithm like Google’s search results. Instead, vendor mentions are influenced by:
- How commonly they appear in publicly available data.
- Whether the AI is prompted to search for them (via a web query).
- Brand awareness and market dominance (e.g., ChatGPT is more likely to mention AWS for cloud services than a smaller provider because AWS is widely known).
🚫 Paid Placement Doesn’t Exist – Unlike Google Ads or paid search, you can’t buy placement or “optimize” your vendor’s presence directly within ChatGPT’s responses.
What You Can Do to Increase Visibility in AI Tools
💡 Brand Awareness & Thought Leadership
✅ Publish authoritative content on industry-specific topics.
✅ Get cited in industry reports, reputable blogs, and news articles.
✅ Ensure your website has clear, structured content about your offerings.
💡 Consistent Mentions in Relevant Contexts
✅ Be active on industry forums, LinkedIn, and expert Q&A sites like Quora or Reddit.
✅ Encourage discussions about your vendor’s unique strengths.
💡 Strong PR & News Coverage
✅ Press releases, case studies, and partnerships can help get your company mentioned in sources that AI models reference when searching the web.
💡 Optimized Website & Knowledge Graph Presence
✅ Ensure your brand is well-represented on Google’s Knowledge Graph and Wikipedia.
✅ Use structured data (schema markup) on your website to improve how AI tools interpret your company’s details.
💡 User Engagement & Direct Interaction
✅ If customers and prospects frequently ask AI tools about your vendor or use your brand name in queries, this could increase recognition over time.
Key Takeaway: You Can’t “SEO” ChatGPT, But You Can Influence It
Unlike traditional SEO, AI discoverability is driven by brand presence across reputable sources. The more analyst reports, media coverage, and authoritative mentions your brand has, the more likely AI tools like ChatGPT and Gemini will reference your company in responses.
If your SaaS brand isn’t appearing in AI-driven research yet, focus on thought leadership, media visibility, and structured data—because that’s how AI finds you.
