WHAT AI SYSTEMS ACTUALLY USE TO DECIDE WHO TO RECOMMEND

AI systems do not recommend businesses based on their websites. They do not look at your homepage headline, your service page copy, or your blog post keyword density. The signals AI uses to decide who to recommend come from somewhere most business owners have never thought about. Understanding those sources is the first step to becoming visible.

The Business Visibility Index studied where AI actually pulls its information from. The answer surprised us.

WHAT AI IGNORES

Your own website, while important for humans, is one of the weakest signals for AI recommendations. AI systems know that you control what your website says. Anyone can write "we are the #1 company in our industry" on their homepage. AI discounts self-reported claims heavily.

AI also largely ignores aggregator sites, people-finder databases, and low-quality directory listings. These sources lack editorial judgment. They collect information without verifying it.

WHAT AI TRUSTS

The sources AI systems weight most heavily share one characteristic: editorial curation. Someone, whether a journalist, a podcast host, a conference organizer, or a professional association, made a judgment call to include you. That judgment is the signal.

Press coverage and publications. When Forbes, Entrepreneur Magazine, or a trade publication writes about you or publishes your content, that is a strong signal. It means a publication with editorial standards decided you were worth featuring.

Podcast appearances. Every podcast episode creates a page with your name, your title, your company, and a description of your expertise. That page lives on Apple Podcasts, Spotify, and the show's website. Three corroboration points from a single appearance.

Professional associations. Being listed on ama.org, your industry association, or a professional directory with membership requirements signals credibility that AI can verify.

Conference speaking. When a conference puts your name on their schedule page alongside other known speakers, AI reads that as peer-level recognition.

LinkedIn activity. Consistent posting, engagement, and a well-structured profile give AI a current, ongoing signal that you are active in your field. A dormant LinkedIn profile tells AI nothing.

Structured data on your website. While your website copy is a weak signal, structured data (JSON-LD schema) is a strong one. Schema is not marketing copy. It is a machine-readable declaration of facts: who you are, what you do, where you have been published, what organizations you belong to. AI reads schema like a filing system, not like an advertisement.

THE CORROBORATION PRINCIPLE

No single source is enough. What matters is corroboration across multiple independent sources. If your LinkedIn says you are a VP at Company X, your Forbes author page says the same, a podcast episode description confirms it, and your website schema declares it, AI has four independent signals pointing to the same fact. It crosses the confidence threshold and includes you in its answers.

If only your website makes the claim and no third-party source confirms it, AI treats it as unverified. You stay out of the answer.

THE ENTITY PAGE TIES IT ALL TOGETHER

The entity page on your website is not about making claims. It is about pointing AI to the independent sources that verify your claims. Every outbound link to a Forbes article, a podcast episode, a conference schedule, or a professional association page is a corroboration signal. The entity page is the hub. The third-party sources are the spokes.

The business owners in the top 2.5% of the BVI (Authority tier) did not get there by having better websites. They got there by having more corroboration. More sources saying the same thing about them, independently, consistently, and verifiably.