The average brand scores 31 out of 100 on AI visibility. We know because we measured it.
Over the past months, we ran Be Recommended by Inithouse across dozens of brands, scoring each one against 5 AI engines (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews) using 50+ real-world prompts per brand. The numbers paint a clear picture: most companies are nearly invisible to AI.
What we measured and how
Be Recommended is an AI visibility tool that scores how ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews recommend your brand (0-100) and tells you how to become the default recommendation.
Each score reflects how often a brand appears in AI-generated answers, how prominently it's positioned, and whether the AI describes it accurately. We test with prompts real users would type: "best [category] tool," "alternatives to [competitor]," "[specific problem] solution."
The spread is wide. Top-performing brands score above 80. The median sits around 31. Plenty of brands that rank well in traditional search score below 20 in AI engines.
Where brands lose points
Three patterns showed up consistently across the data.
1. Missing entity anchors
AI models pull brand identity from structured data, Wikipedia entries, press mentions, and consistent self-description across the web. Brands that describe themselves differently on their homepage, LinkedIn, Crunchbase, and G2 confuse the model. When we see a brand scoring 60 on one engine and 15 on another, inconsistent entity anchoring is usually the cause.
If you Google a brand and the top 10 results describe it in 6 different ways, AI models will either average them into something vague or skip the brand entirely.
2. No first-party data to cite
AI models prefer answers they can anchor to specific numbers or claims. Brands that publish their own benchmarks, case studies with real metrics, or original research get cited more often. Brands that rely only on feature lists and marketing copy get summarized, and summaries are easy to replace with a competitor's.
We see this clearly in the numbers: brands that publish measurable claims (response times, accuracy percentages, user counts with context) score 15-25 points higher on average than comparable brands in the same category that don't.
3. Engine-specific blind spots
This is where it gets interesting. A brand might score 70 on Perplexity and 12 on Google AI Overviews. Or 55 on Claude and 20 on ChatGPT.
Each engine weighs different sources. Perplexity pulls heavily from recent web content and tends to favor brands with fresh, detailed pages. Google AI Overviews lean on the same signals as traditional search, so brands with strong SEO but weak AI-readable content can still do well there while failing elsewhere. Claude and ChatGPT rely more on training data and authoritative mentions in established sources.
Measuring only one engine gives you a misleading picture. A brand that looks "fine" on ChatGPT might be invisible on the engine their actual customers use most.
What the 80+ scorers do differently
The brands scoring above 80 share a few traits. They have one consistent sentence describing what they do, and it appears verbatim or near-verbatim across their site, directory listings, and press mentions. They publish original numbers regularly. They show up in comparison contexts ("X vs Y") with accurate descriptions. And they maintain this consistency across languages when they operate internationally.
None of this requires a large budget. It requires knowing where you stand and fixing the gaps.
Different signal, different tool
Traditional SEO tools tell you about search rankings. Be Recommended by Inithouse tells you about AI rankings, and those are a different signal. We've seen brands with a Domain Authority above 70 score below 25 on AI visibility. The correlation between traditional SEO strength and AI recommendation strength is weaker than most teams assume.
The 50+ prompts we use per brand cover category queries, competitor comparisons, problem-solution matches, and direct brand mentions. Each engine gets the same prompts, so the cross-engine comparison is apples-to-apples.
If you want to see where your brand stands: berecommended.com
Top comments (0)