G2 has recently launched G2.ai, an “AI driven discovery” bolt on that sits awkwardly on top of its once dominant software marketplace. It promises faster conclusions, smarter suggestions and personalised shortlists. That all sounds lovely, until you remember one awkward fact: we can already get that – and more – from the AI assistants living in the tabs we keep open all day. G2 is trying to claim a piece of an AI search future that has already rushed past it.
The numbers do not lie: G2’s traffic is in retreat
Public traffic estimates from tools like Similarweb show G2’s monthly visits sliding. While the exact figures vary depending on the data source, the direction of travel is brutally clear: fewer people are actually using legacy reviews/comparison marketplaces to compare software.
Meanwhile Google – the gateway G2 still relies on for visibility and referral traffic – has started injecting AI Overviews directly into results. Ask a buying question like “best CRM for SMB” and you will increasingly see a generated answer that summaries multiple sources, pushes the organic listings halfway down the page and removes the need to click through. Every time that happens, G2 bleeds a little more search driven oxygen.
AI search tools have already eaten G2’s lunch (and data)
G2.ai frames itself as a revolutionary layer, but in reality it is feeding on the same corpus the wider AI world hoovered up months ago. Crawl restrictions were too little, too late. Every major model in market has already ingested millions of G2 reviews, comparison tables and category descriptions. Whether or not that wholesale scraping was ethical or legal is a fascinating legal sub plot – but it is has already happened, whether you like it or not.
The result is simple: when I ask ChatGPT, Claude or Perplexity for a software recommendation, they reply with a blended, perspective rich answer that references G2, Capterra, Gartner Magic Quadrants, Forrester Wave reports, analyst commentary and user forums. G2.ai can only echo one of those sources – itself.
Narrow data, narrow answers
G2.ai insists it provides a “trusted, expert” shortlist. Let us decode that in the context of AI search: it is a shortlist constrained by a single data silo. Even if you believe every G2 review is squeaky clean and free of vendor gaming (unlikely), the most it can offer is a statistical view of its own walled garden. The wider context – total cost of ownership analysis, integration pitfalls, regional support quality, roadmap credibility – lives outside G2’s perimeter, and therefore outside G2.ai’s reasoning.
Contrast that with your personal AI assistant of choice. It remembers the systems you have already ruled out, the tools you already use, the budget ceiling finance slammed on your last project, the tech stack your architects refuse to touch and even your colleague in IT security that doesn’t want any more software tools. That context shapes the next recommendation instantly. G2.ai cannot do that because it does not know you, and because its creators are still dragging legacy review workflows into an age of conversational search.
A patch on a leaking hull
G2 is not the first incumbent to slap an AI sticker on its product in hope investors can breathe easier. The problem is deeper than branding. Marketplaces that rely on capturing search intent between keyword and vendor website are being disintermediated. AI search answers that intent directly and instantly, inside the viewport where the question was asked. Adding a chat search box on G2’s own site does nothing to reverse that macro shift.
G2.ai is, at best, an undersized plug jammed into the hole below the waterline. It might slow the flooding briefly. It will not pump the water back out, and it definitely will not stop the ocean creating new holes tomorrow.
What does it mean for B2B marketers?
G2 can still be a highly valuable source for visibility. But not for the reason that it used to be.
LLMs love training on review data from sites like G2. We know from lots of our own research that G2 often will be referenced by AI search tools like ChatGPT as an influential source for a B2B software search.
So the value of being on G2 is now that it indirectly plays into your AI search visibility strategy. A strong G2 profile may result in more likelihood of being referenced by a AI tool for a search relevant to your category. So it’s still worth investing in your G2 presence – not for direct clicks and referrals, but to optimise for AI search (AEO, LLMO, AIO, GEO – whatever the process of optimising for visibility in AI search is actually called).
Final thought: An AI Band-Aid on a Sinking Ship
We are not writing G2’s obituary – yet. There is still a role for independent reviews of B2B software – but the landscape has changed massively. But for now, it seems G2 is trying to claim a piece of an AI search future that in many respects has already rushed past it.