AI visibility, explained.

For vendor discovery, yes. 24% of B2B buyers now use AI as their primary research tool. That number was 2% in 2024 (GrowthSpree, 2026). Google still handles most searches. But the searches that matter — what tools do teams like mine use for X — are moving to ChatGPT, Claude, and Perplexity faster than most companies realize. By the time a buyer reaches your website, the shortlist is often already built.

94% used an LLM at some point during their most recent purchase (6sense, 2025, n=4,510). 24% start there. 67% conduct research without filling a single lead-capture form, up from 41% in 2024 (GrowthSpree, 2026). Most of that research happens inside AI tools. Vendors who are not visible at that stage are not being evaluated.

Buyers arrive at vendor conversations with 2 to 4 options already chosen. That list was built by AI before they spoke to anyone. Before AI, buyers started with 7 to 12 options and narrowed down through research. Now they start narrow. The pre-contact favorite wins 80% of the time (6sense, 2025). If you are not on the Day One list, you almost certainly will not win the deal.

They convert at 14.2 to 16.8% — Claude referrals top that range. Google organic converts at 1.76% (Seer Interactive, 2026). AI visitors spend 68% more time on site and bounce 33% less (SE Ranking, 2025). They arrive already trusting the recommendation. It is the highest-quality traffic available in B2B, and most companies are not generating any of it.

ChatGPT leads at 62.6% of AI referrals. Claude is second at 18.5% and growing fast — it moved from 1.4% to 18.5% in 8 months. Gemini holds 10.6%, Perplexity 7.3%, Copilot around 4% (HiGoodie Wave 2, March–April 2026). Claude is now the fastest-growing B2B referral source in AI.

68% of all Google searches end without a click (SparkToro / Similarweb, April 2026). When AI Overviews appear, that number reaches 83% (Bain/Dynata, Dec 2024). Organic CTR fell 61% on queries triggering AI Overviews (Seer Interactive, 2.43 billion impressions). Traffic from Google is not just declining in quality. Volume is falling too.

Three factors dominate. First, AI crawler access: if your robots.txt blocks named AI crawlers, those systems cannot index you. Second, third-party earned media: 85% of brand mentions in AI come from external sources, not your own site (AirOps, 21,311 mentions). Third, entity clarity: consistent facts across LinkedIn, Crunchbase, and press materials reduce AI hallucination about who you are. Content on your own domain is a distant fourth.

No. 80% of ChatGPT, Perplexity, and Copilot citations do not rank in Google's top 100 (AIVO, March 2026). Only 38% of Google AI Overview citations appear in the organic top 10 (Ahrefs, March 2026, 863K keywords). Web mentions correlate with AI visibility at r=0.664. Backlinks correlate at r=0.218 (Ahrefs, Dec 2025, 75K brands). The signals are different. Optimizing for Google does not optimize for AI.

No. A controlled Ahrefs study published in May 2026 analyzed 1,885 URLs. Adding schema markup produced zero causal impact on AI citation rates for pages already achieving AI visibility. Schema serves Google rich results. It does not influence how AI systems retrieve and cite content. Do not spend time on it for AI visibility purposes.

No. Ahrefs analyzed 137,000 domains in May 2026 and found 97% of live llms.txt implementations receive zero AI crawler requests. Google does not read it. Only 1 URL in 94,614 AI-cited URLs pointed to an llms.txt file. It is not a lever. Treating it as one wastes attention that should go to things that actually move the needle.

robots.txt. Seven named AI crawlers are blockable: ClaudeBot (Anthropic), GPTBot (OpenAI), PerplexityBot (Perplexity), Google-Extended, Applebot-Extended (Apple), CCBot (Common Crawl), and cohere-ai (Cohere). Many sites block them through broad wildcard rules without realizing it. Unblocking takes five minutes. It is the single highest-leverage action available with no downside.

Dense marketing prose fails AI parsers. PDFs fail. JavaScript-rendered content that only displays after page load fails. Clear HTML headings, visible structured content, and direct declarative statements work. When AI cannot parse your content, it fills the gaps with fabricated averages about your category. The fix is structural, not stylistic.

84% of all AI citations are earned media: journalism, analyst coverage, community mentions (Muck Rack, May 2026, 25M links). Paid content accounts for 0.3%. Your own domain accounts for 13.2%. Brands are 6.5x more likely to be cited via earned media than their own domain (Position Digital, 2026). PR and unlinked brand mentions outperform any on-site optimization you do by a wide margin.

Domain in, automated analysis out. You submit your domain. We run analysis across five dimensions and return a complete picture of your AI footprint: what AI systems say about you, who they recommend instead, and where the gaps are. One briefing, structured output, prioritized next steps.

Retrieval Access: your robots.txt checked against 7 named AI crawlers. AI Narrative: 20 search-grounded model calls with verbatim responses captured. Competitive Position: competitors discovered from AI responses only, never self-reported. Citation Source Map: your top 5 citation sources, depth-scored, gaps quantified. Entity Clarity: Crunchbase, LinkedIn, Wikipedia, and live knowledge panel queries cross-referenced against your own claims.

Self-reported competitors corrupt the signal. Most companies self-report based on who they know, who they have lost deals to, or who their sales team watches. AI does not share that frame. It builds competitive sets from training data and real user behavior. If you ask us to benchmark against a competitor AI never mentions, you have measured the wrong thing. We discover competitors the way your buyers discover them.

B2B SaaS companies with an existing product and some market presence. The analysis is most useful when there is a gap between how you describe yourself and how AI describes you. That is true for 89% of B2B companies (MaximusLabs, 2026, n=200+). If you are pre-launch with no domain history, wait until you have something for AI to retrieve.

The product is in development. Delivery time will be published at launch. Join the early access list to be notified first and get weekly AI visibility insights while we build.

Price is open. We are testing at launch. The first 100 completions are free. Join the list to be in that window.

Ask ChatGPT, Claude, and Perplexity the questions your buyers ask: what tools do teams use for [your category], and compare the top vendors in [your space]. If your brand does not appear, or appears with wrong facts, you have an AI visibility problem. 62% of enterprise brands are invisible to AI despite 94% investing in traditional SEO (ALM Corp, 2026, n=1,000). Most companies do not check until a competitor is already owning the answer.

robots.txt. Check whether ClaudeBot, GPTBot, PerplexityBot, Google-Extended, Applebot-Extended, CCBot, and cohere-ai are blocked. If any are, unblock them. It takes five minutes and it is the highest-leverage single action available. After that: PR and third-party coverage. Everything else is noise until those two are addressed.

The practice of structuring your brand so AI systems can find, understand, and cite you. It includes crawler access, entity consistency, content structure, and earned media strategy. It is distinct from traditional SEO because the ranking signals, sources, and effective content formats are all different. The Agent Native briefing measures your current baseline across all five dimensions.

The newsletter covers one AI visibility insight per week: what is moving, what the data shows, what to do. The briefing gives you a complete audit of your specific domain. Start with the newsletter to understand the space. Get the briefing when you are ready to act on it.