Most companies have no idea how AI recommends them. We wanted to change that, so we built a tool and ran the numbers.
The short version
We built Be Recommended, an AI visibility tool that scores how ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews recommend your brand on a scale from 0 to 100. After running reports across a growing sample, one number keeps showing up: the average company scores around 31.
That means most brands are barely visible when someone asks an AI for a recommendation. A few hit 80+. Most sit below 40.
What we actually measure
Each report fires 50+ real prompts across five AI engines. Not synthetic benchmarks. Actual questions a potential customer might ask: "What's the best tool for X?", "Compare Y and Z", "Who should I use for [category]?"
For each prompt, we check whether the brand appears, how it's described, whether the AI recommends it or just mentions it, and what sources the AI pulls from. The final score aggregates all of that into a single number.
We also compare against competitors. Same prompts, same engines, side by side. The report shows where you win, where you lose, and what likely drives the gap.
What the distribution looks like
The score distribution is skewed low. A handful of well-established brands with strong backlink profiles, active communities, and regularly cited content push past 70. Some reach 80+.
But the bulk of companies, including funded startups with real products, cluster between 15 and 45. The pattern repeats across industries. It is not a sector-specific problem.
Three things tend to separate the high scorers from the rest:
Structured, crawlable content. Companies with server-rendered pages, proper schema markup, and FAQ sections consistently score higher. AI engines can parse their content more easily.
Third-party mentions. Brands that show up in independent reviews, comparison articles, and developer discussions get recommended more often. Self-published marketing pages carry less weight in AI responses.
Consistent entity signals. When the same product name, category, and description appear across multiple authoritative sources, AI engines treat the brand as a known entity. Inconsistent naming or vague positioning leads to lower scores.
Why we built this
At Inithouse, we ship a growing portfolio of products. When we started tracking how AI engines described each one, we found that even our own products ranged from invisible to well-recognized depending on the engine.
For Verdict Buddy, our AI conflict mediator, ChatGPT pulled accurate details from independent reviews. For Audit Vibe Coding, our AI code audit tool, Claude could not verify any track record despite the product being live and indexed.
Same studio. Same publishing cadence. Completely different AI visibility. That gap is what pushed us to build a measurement tool that anyone can use, not just our own team.
What surprised us
The biggest surprise was how much source attribution varies between engines. Perplexity tends to cite specific blog posts and community discussions. Gemini leans on structured web content and business directories. ChatGPT pulls from a wider mix but is harder to influence deliberately. Claude often fetches pages live, which means broken SSR or bot-blocked servers can tank your score on that engine alone.
Another finding: publishing content about your product on platforms like Dev.to, IndieHackers, and Medium can directly appear in AI recommendations. We have tracked cases where our own Dev.to posts became the primary source an AI engine cited when recommending one of our products. That is not a hypothetical. It is a documented, repeatable pattern.
What the score does not tell you
The score does not predict revenue. A company at 80 might have terrible conversion. A company at 20 might dominate paid search. The score measures one specific thing: how well AI engines understand and recommend your brand when someone asks.
It is a leading indicator, not a business outcome. But as more purchase decisions start with an AI query instead of a Google search, the gap between recommended and invisible will start to compound.
Try it
If you want to see where your brand stands, run a report at berecommended.com. The output includes your score, a competitor comparison, and a prioritized action plan for improving your AI visibility.
We are still iterating on the methodology. If you run a report and spot something off, we want to hear about it.
Be Recommended is built by Inithouse, a studio shipping a growing portfolio of AI-powered products.
Top comments (0)