When a potential buyer wants to evaluate your software product or service offering today, there’s a good chance they don’t start with Google. They open ChatGPT or use Google’s AI Mode and ask something like: “What’s the best project management tool for a 50-person team?” or “Is [your brand] worth it?”
What they get back isn’t a list of links to relevant websites as is the norm for a Google or Bing search. Rather, they get a verdict about the best option for their perceived needs. It’s a characterisation assembled from dozens of sources your marketing team has never audited, shaped by signals you’ve probably never thought to manage. It’s the embodiment of AI brand reputation exposure in action.
This is a significant change that most software as a service (SaaS) marketing teams haven’t fully reckoned with yet. AI hasn’t just become a new discovery channel. It’s become your brand’s editor – compressing your entire digital footprint into a single response, forming an opinion about you, and publishing that opinion whenever it mentions your product. And it didn’t ask your permission.
The new now – AI isn’t neutral
For two decades, search worked a particular way. Google returned a ranked list of pages. The searcher clicked around, formed their own view, and made a decision. The search engine was a librarian – it pointed you toward sources, but it didn’t tell you what to read or think.
That’s no longer how it works.
AI search engines don’t just retrieve information – they interpret it. They weigh sources, identify patterns, synthesise competing signals, and deliver a single coherent response that carries the implicit authority of the engine producing it. For many people, that response is the research.
Recent research has illustrated how this works. In March 2026, BrightEdge published its findings after analysing hundreds of millions of prompts across Google AI Overviews and ChatGPT, tracking every brand mention and its sentiment. The stats about how frequently negative sentiment is surfaced in brand mentions is striking.
- Google AI Overviews: 2.3% of brand mentions
- ChatGPT: approximately 1.6% of brand mentions
Google is 44% more likely than ChatGPT to characterise a brand negatively overall. But the more revealing finding was how each engine goes negative – because they do it very differently.
Google AI Search behave like an investigative reporter. When they go negative, it’s typically triggered by controversy: lawsuits, data breaches, regulatory actions, product recalls, boycotts. If your brand has ever had a public incident of this kind, Google’s AI engine is likely to surface it – repeatedly, to everyone whose query touches your category.
ChatGPT, by contrast, behaves more like a product advisor. Its negative characterisations tend to focus on feature limitations, pricing concerns, compatibility issues, and the kind of evaluative questions – “Is it worth it?”, “What are the downsides?” – that buyers ask when they’re close to a decision.
The same brand, on the same query, can be treated positively by one AI search engine and negatively by the other. And when BrightEdge analysed overlapping prompts where both engines surfaced negative sentiment, they flagged different brands 73% of the time – despite responding to identical queries.
These percentages may sound manageable. They’re not. Across billions of monthly interactions, 2–3% of negative characterisation translates to millions of brand-negative exposures, served at scale. They’re included in the AI-generated summary rather than (hopefully) being hidden on page-two of traditional search results.
The stakes are higher for SaaS brands
Every industry faces this shift, but SaaS brands are particularly exposed – and for reasons that are structural rather than incidental.
B2B SaaS buyers are among the most research-intensive buyers in any market. Before they talk to sales, they’ve typically already read reviews, compared alternatives, checked community discussions, and formed a strong preliminary view. That research journey (which used to unfold across multiple platforms over days or weeks) is now being compressed into a handful of AI queries.
The other factor that raises the stakes for SaaS is where ChatGPT’s negativity tends to land. BrightEdge found that while Google AI Overviews concentrate 85% of their negative sentiment at the informational stage – the research and discovery phase – ChatGPT surfaces 19.4% of its negative mentions during the consideration-to-purchase phase. That’s 13 times higher than Google’s 1.5% at the same stage.
In practical terms: Google’s negative characterisations tend to influence the top of the funnel. ChatGPT’s tend to kill conversions near the point of purchase.
For SaaS brands with longer sales cycles, this is a pipeline risk that doesn’t show up obviously in your analytics. A prospect who was nearly ready to convert, who asked ChatGPT one final comparative question and got a characterisation that raised doubts – that lost opportunity is invisible. It just looks like a deal that went cold.
Where AI gets its opinions about your brand
The question SaaS marketing teams need to ask isn’t just “what is AI saying about us?”, it’s “what is AI reading to form that view?”
The answer is more varied, and more difficult to control, than most teams assume.
Review platforms carry enormous weight. G2, Capterra, and TrustRadius are among the most heavily cited sources in AI-generated responses about software. Research by SE Ranking found that review platforms appear in roughly one third of AI Overviews, with G2, Capterra, Gartner Peer Insights, Software Advice, and TrustRadius accounting for 88% of all review platform citations in those responses. For SaaS brands, presence on these platforms isn’t optional – but presence alone isn’t the whole story.
Recency matters more than volume. AI engines, like consumers, give more weight to recent content. Research from Bright Local found that 74% of consumers only trust reviews from the last three months, and AI systems reflect the same preference. A review profile dominated by two-year-old reviews, even overwhelmingly positive ones, can result in AI characterisations that feel outdated.
Reddit and community forums feed AI training data. AI models are trained on, and retrieve from, a vast body of community-generated content – and that content doesn’t expire the way a news story does. A Reddit thread from 2022 complaining about your onboarding process, or a comparison thread where people enthusiastically recommended a competitor, can still be influencing how AI characterises your brand today.
Your own content has some sway. AI engines do draw from your website, but they cross-reference your marketing claims against third-party sources. If your site positions your support as fast and fully informed, but a cluster of recent reviews describe it as frustrating, AI tends to weight the independent signals more heavily. The inconsistency itself becomes a risk factor.
In practical terms, this means that many sources have to be monitored. One of the most underappreciated sources of reputation drag in the current AI search environment is old community sentiment that was never countered with fresh, accurate information. AI engines don’t check whether the problem being discussed has been fixed. They read the content as written and factor it in. It’s important to keep third party sources of information up to date.
Likewise, a burst of review collection once a year is far less effective than a steady, ongoing cadence that keeps your profile current and representative of where your product actually is today, not where it was eighteen months ago.
Why traditional tactics fall short
Understanding where AI gets its signals is useful. But knowing what to do about it starts with understanding where a conventional approach falls short.
Most SaaS marketing teams are running a strategy that was designed for a fundamentally different version of search. Traditional search engine optimisation (SEO) focused on ranking your own pages for target keywords – optimise the page, build authority, climb the search engine results page (SERP). It was primarily about ensuring Google could read your site and your intent properly.
That playbook still matters. But it was never designed for a world where AI engines synthesise the entire web – including content you don’t own and conversations you’re not part of – into a single editorial response about your brand.
Traditional SEO doesn’t tell you what Reddit is saying about your product. It doesn’t help you manage the review freshness that AI engines use to calibrate their characterisations. It doesn’t account for the fact that G2 and Capterra, despite losing up to 92% of their organic traffic from traditional search, are now cited more frequently inside AI-generated answers than almost any other source. And it doesn’t give you any visibility into what ChatGPT is telling buyers who are minutes away from a purchase decision.
The gap isn’t in the fundamentals of good content or technical SEO – those remain relevant. The gap is in scope. A strategy that only optimises for what you control is increasingly insufficient in a search environment shaped heavily by what you don’t.
Five things to do, starting today
None of this requires a complete reinvention of your marketing approach. But it does require expanding where you focus your attention and adding some disciplines that most SaaS marketing teams haven’t formalised yet.
- Audit your AI presence before you do anything else.
Run consistent prompts across Google AI Overviews, ChatGPT and Perplexity, using the same language your buyers actually use. What comes back? Is the characterisation accurate? Is it pulling from outdated information? Does it reflect your current product, or your product from two years ago? This audit should happen quarterly, not once. AI responses change as the web changes, and your monitoring needs to keep pace. - Treat review platforms as AI infrastructure.
G2 and Capterra are no longer just lead-gen channels – they’re primary inputs into what AI says about your brand. Your goal is to maintain a steady flow of recent, substantive reviews that give AI engines current, representative signals. An automated cadence that prompts satisfied customers to leave reviews at natural moments in the customer lifecycle will outperform any one-time review campaign. - Address the recency problem in community discussions.
You can’t delete old Reddit threads, but you can counter them with fresh, accurate content in the same spaces. If there are persistent community discussions about issues your product has resolved, being present in those conversations – or creating new content that addresses those questions directly and accurately – gives AI engines more recent signals to draw from. The internet’s default position is silence; you have to actively fill the gap with current information. - Align your content strategy with how AI reads the web.
Fresh, well-structured content that AI can parse and attribute clearly is more likely to be drawn upon in AI responses. This means thinking about your content not just in terms of ranking potential, but in terms of what characterisation it contributes to your brand’s overall AI footprint.
Structured comparison content, transparent feature documentation and clear positioning against alternatives all help AI engines form a more accurate picture of what you offer and for whom. This is where search strategy and content strategy – optimised for the generative engine optimisation (GEO) era – take the lead. - Monitor both Google and ChatGPT as separate environments.
They draw from different source ecosystems and go negative for different reasons, so a monitoring approach that only watches one platform limits effectiveness. Google’s negativity tends to be controversy-driven in a way that impacts early searchers. ChatGPT’s tends to be evaluation-driven and conversion-killing. Each requires its own awareness and its own response.
The shift has already happened
AI has assumed editorial authority over your brand. It’s not a future state to prepare for – it’s the environment your buyers are already navigating.
For SaaS brands, where trust is foundational and the buying journey is research-intensive, this has real consequences for pipeline, conversion, and competitive positioning. The brands that adapt are those that understand the characterization AI constructs from the full body of signals it reads. It’s those brands that will benefit from a search and content strategy that evolves to reflect the environment.
What AI says about your brand is already influencing your buyers. The question is whether you’re paying attention.
If you’re not sure what AI is currently saying about your brand, or whether your search and content strategy is equipped for this shift, we’d be glad to talk it through. Or if you’d like to dig deeper into the mechanics, our Engage AI SEO Playbook is a good place to start.
