Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

The shift from SEO to GEO

Four ways SaaS leaders can compete for visibility in AI-driven answers.

[Photo: Getty Images]

Profile photo of Anton Storozhuk

President · Metrica Software Inc

A good product and a user-friendly website were once enough to drive customers from search. Today, potential customers increasingly ask artificial intelligence assistants what service to choose before even opening Google. If the brand isn’t in the AI-generated answer, the client may never make it to the website.

A year after AI Overviews launched, the share of Google searches ending without a single click to a website has grown from 56% to 69%. A survey of B2B professionals found that AI already shapes 92% of vendor shortlists and 83% of final vendor purchase decisions. Brand recognition determines shortlisting in only 7% of cases. A precise match to the customer’s request matters far more. AI still favors established names within categories, but in 93% of cases, precision beats recognition.

What wins is how closely a brand matches the request. That’s the opportunity for a lesser-known brand to outmaneuver a market leader.

HOW AI CHANGED THE SEARCH RULES

Search engine optimization and generative engine optimization solve the same problem in different ways: How does an algorithm decide what to show in response to a query? SEO matched keywords, links, and meta tags on a page to produce a list of websites. With GEO, AI systems read many sources on a topic to generate a single, coherent answer.

Backlinks and brand mentions remain important for both SEO and GEO. What changed is what the algorithm counts as authoritative. Before citing a fact, an AI assistant verifies it across multiple independent sources.

Volume used to be the most important criterion: The more backlinks led to a website, the higher its position was. That made the system easy to trick with link farming. A website could hold a top spot on Google for years purely due to a huge backlink volume, even if almost no media outlets ever mentioned the company.

AI systems work differently. The number of mentions doesn’t matter. The authority of sources where these mentions appear and their consistency do. AI won’t perceive a website backed by a thousand backlinks but with no real media coverage as a credible source.

The data confirms this: If you rank a hundred brands by backlink count and then by visibility in AI Overviews, these two lists will barely overlap. Rank them by mentions in independent sources instead, and the lists will line up far more closely.

Since AI assistants rely on content and consistency of mentions, tackle GEO priorities in the order below.

1. SEE WHAT AI IS ALREADY SAYING ABOUT YOUR BRAND

First, ask an AI assistant the same questions a customer would and see which sources it cites.

A one-off check won’t give you the full picture because answers can change daily depending on the way you phrase the same question. To skip the manual work, I use Brand24 that tracks brand mentions in ChatGPT, Gemini, and Perplexity answers automatically and shows share of voice against competitors.

2. MAKE YOUR SITE ACCESSIBLE TO CRAWLERS AND PUT THE ANSWERS UP FRONT

To earn AI citations, meet two conditions: Bots need access to your website, and key facts must be immediately visible.

First, never block bots like GPTBot and ClaudeBot in robots.txt. If a crawler can’t reach a page, nothing will help, even adding schema markup. Second, 44.2% of citations come from the first 30% of a page. Key takeaways belong at the top.

I use Yoast SEO to check if my website’s pages meet both conditions. It flags whether the crawler is blocked and the heading hierarchy is correct. For internal linking I use InLinks, which helps AI systems find the relevant content on my website faster.

3. GET YOUR CONTENT REGULARLY PUBLISHED IN INDEPENDENT MEDIA

According to Muck Rack, 82% of all AI citations come from earned media—independent coverage, not a company’s own content.

We regularly publish content about our company in relevant media outlets. As negotiating with individual outlets takes time and AI assistants don’t distinguish between earned and paid media, I use PRNEWS and Medialister. They provide access to a pool of media outlets at once. Coverage runs faster than direct negotiations.

4. MANAGE YOUR REPUTATION ON REVIEW PLATFORMS

AI assistants quote reviews verbatim. ChatGPT, for instance, will phrase an answer like “The company’s rating on Trustpilot is …” AI also absorbs the emotional tone of reviews. If a brand has mostly negative reviews, the model’s answer will feel negative too, and vice versa for positive reviews.

Review volume and recency matter too. A profile with five three-year-old reviews and an actively updated profile with responses from the company are perceived differently. AI assistants will treat the latter as a more credible source. That’s why I ask clients to leave reviews on Trustpilot and G2. I also track brand mentions in relevant communities on Reddit. It’s less predictable than reviews on dedicated platforms, but an organic recommendation there is a strong signal for AI.

None of these four steps delivers results on its own: A website with no mentions in media outlets won’t be cited by AI. Mentions without reviews won’t set the right tone.

WHY WAITING COSTS MORE

The broader and more established the category, the harder it is to break into that fixed list from scratch. A company should define its niche precisely. Instead of “CRM” use “CRM for distributed B2B sales teams.”

A client once told me he’d searched for a customer relationship management system for his remote sales team in ChatGPT, skipping Google completely. Only three brands made it into the answer—and he had never even opened any other websites. Companies that didn’t make the list lost the deal before they ever knew the client existed. If you’re not investing in GEO, you may have already lost hundreds of deals.

Anton Storzhuk is CEO of Metria Software.