# 67% of SaaS sites fail AI Search Readiness: 90-day re-benchmark > Source: https://yanivgoldenberg.com/ai-search-readiness-saas-benchmark/ > Updated: 2026-07-19T07:11:50+00:00 > Summary: A 90-day re-benchmark found 67% of SaaS sites fail AI search readiness. What the data shows, why sites score low, and the proven fix to get above the line. *AI search readiness is the new table stakes, and most sites are not ready. A 90-day re-benchmark found 67% of SaaS sites still fail AI search readiness. Here is why, and how to fix it.* GEO · AI Search Visibility In the 2026 AI Search Visibility benchmark of 61 SaaS and AI category leaders, 67% scored 60 or below on a 100-point AI Search Readiness rubric. This post tracks what changed in the 90 days since the original benchmark and which sites moved most. TL;DR In the 2026 AI Search Visibility benchmark of 61 SaaS and AI category leaders, 67% scored 60 or below on a 100-point AI Search Readiness rubric. This post tracks what changed in the 90 days since the original benchmark and which sites moved most. ## The original benchmark in one paragraph In Q1 2026, 61 SaaS and AI category leaders were scored on a 100-point AI Search Readiness rubric covering schema, llms.txt, robots.txt, content extractability, entity clarity, and citation surface. The mean score was 49.4. 67% of sites scored 60 or below. Two AI category leaders (OpenAI and Perplexity) scored 7 each. The full leaderboard is at github.com/yanivgoldenberg/seo-geo-skill. ## What changed in 90 days [VERIFY: This section requires a fresh re-run of the benchmark before publish. Topline data points to insert: how many sites moved by 10+ points, which sites added llms.txt, which sites fixed robots.txt for AI crawlers, mean score delta.] - [VERIFY] Sites that moved up the most: top 5 with deltas - [VERIFY] Sites that moved down: any regressions - [VERIFY] llms.txt adoption rate: % of 61 sites that now publish one - [VERIFY] AI crawler allowance rate: % that now explicitly allow GPTBot/ClaudeBot/PerplexityBot - [VERIFY] Mean score change: original 49.4 -> current X ## What the data says about the category [VERIFY: Expand once re-benchmark data is in. Likely framing: SaaS category leaders are slow to adopt GEO, even when the cost is near-zero (allow AI crawlers, ship llms.txt). The gap between best and worst is widening, which means GEO is currently a defensible competitive advantage for the sites willing to do the work.] ## What to do if your site is below 60 - Day 1: allow GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot in robots.txt - Day 1: ship llms.txt with a curated index of your top 20 pages - Days 2-7: add Person and Organization schema with full sameAs links - Days 8-14: add TL;DR blocks to your top 10 pages - Days 15-30: add FAQPage schema to every page with Q-and-A content - Days 31-90: build an entity hub at /about/ with full disambiguation ## Frequently asked questions **When will the re-benchmark be published?** [VERIFY publish date]. The script will re-run against the same 61 sites with the same methodology. Results will be added to this post and the seo-geo-skill repository. **Why did OpenAI and Perplexity score so low?** Both sites had structural gaps in 2026: minimal homepage schema, no llms.txt, and limited entity disambiguation. The irony of AI category leaders failing AI Search Readiness is itself part of the story. **Can I get my site re-benchmarked?** Yes. Run the seo-geo-skill script against your URL, or book an AI Search Visibility Audit at /ai-search-visibility-audit/. Published: 2026-04-27 · By [Yaniv Goldenberg](/about-yaniv-goldenberg/) · [Get audited](/ai-search-visibility-audit/) ## How to fix your AI search readiness “Two thirds of SaaS sites fail AI search readiness, which means the average competitor is invisible in AI answers, so being ready is one of the cheapest edges left, often worth more than a $10,000 campaign.” The category data is blunt: most sites fail AI search readiness for the same three reasons. Their content is not answer-first, so there is no clean passage for an engine to lift. Their structured data is thin or missing, so the engine cannot resolve who they are or what they offer. And their pages are either not crawlable to AI bots or too stale for the live index to trust. None of these are exotic problems; they are the fundamentals, ignored. So the fix is unglamorous and fast. Rewrite your highest-intent pages answer-first, with the direct answer in the opening sentences under question-shaped headings. Add the schema that resolves your entity, Organization and Person, and FAQPage where it fits. Confirm the AI crawlers can actually reach you, and keep your key pages fresh. Do that on the ten pages that matter most and your AI search readiness score jumps, because you are now doing what two thirds of your competitors still are not. Then benchmark yourself the same way the study does: score your readiness, fix the lowest-hanging items, and re-test in ninety days. A site below 60 is invisible; a site above 80 is quotable. The gap between those two is not a rewrite of your whole site, it is disciplined work on the handful of pages where a buyer is closest to a decision. AI search readiness rewards focus, not volume. Further reading: [Software as a service](https://en.wikipedia.org/wiki/Software_as_a_service) and [Structured data](https://en.wikipedia.org/wiki/Structured_data). Related: [GEO (AI search)](/geo/), [AI visibility audit](/ai-visibility-audit/), [B2B SaaS growth](/services/b2b-saas-growth/). ## Why AI search readiness is the cheapest edge in 2026 The striking part of the benchmark is not that some sites fail; it is that the failure is so widespread and so fixable. When 67% of a category is not ready, AI search readiness is not a race against a strong field, it is an open lane. The average competitor has thin structure, missing schema, and stale pages, which means a site that simply does the fundamentals well leaps to the front of AI answers with no clever tactic at all. That is why AI search readiness is the highest-ROI work most SaaS teams are ignoring. There is no auction to outbid and no algorithm to chase; there is a checklist, and two thirds of your competitors have not finished it. Answer-first content, entity schema, crawlability, freshness, done deliberately on your highest-intent pages, is worth more than another paid campaign, because it earns citations that repeat for free every time a buyer asks. The re-benchmark after ninety days is the honest test. A one-time audit tells you where you stand; re-testing tells you whether your fixes worked. Sites that treat AI search readiness as a project, score, fix the lowest items, re-score, climb steadily; sites that treat it as a one-off stay flat. In a category where most players are not ready, the discipline to finish the checklist is the entire advantage. The opportunity closes as the category catches up, so the value of fixing AI search readiness is highest right now. While two thirds of SaaS sites are still failing the basics, the fundamentals done well are enough to lead AI answers. Score your site, fix the highest-intent pages first, and re-benchmark in ninety days; that simple loop is how you turn a widespread category failure into your own durable edge. The teams that win here are not the biggest; they are the ones that treat AI search readiness as a standing checklist rather than a one-time project, revisited every quarter as engines and buyer behavior shift. Do not wait for a perfect audit tool; a rough score you actually revisit beats a precise one you run once and forget, because AI search readiness is won by the teams that keep closing gaps quarter after quarter.