How does AI decide who to recommend?

Not by magic, and not by accident. This is what happens when an AI assistant answers a customer's question with a business name attached.

When someone asks an AI assistant a question like "who's the best [your industry] near me," it isn't looking up a directory. It's piecing together an answer from whatever it can find written about businesses like yours, then deciding which one or two to actually name.

Where it's getting its information

Usually some mix of the following, in no fixed order:

This is exactly what a Compass North snapshot reports back to you: which of these the AI actually leaned on when it answered a question about your business.

Why two similar businesses get different answers

If the AI can find clear, matching information about one business and only patchy or conflicting information about another, it tends to favor the one it can describe with confidence. That's not a ranking algorithm you can game; it's closer to how a person would recommend a business they could describe clearly over one they only half remember.

Why this isn't the same everywhere

Different AI assistants weigh these sources differently, and the same question asked two different ways can get two different answers. That's why a single "is my business good at this" check doesn't tell you much. A snapshot tests the real range of questions your customers actually ask, across the assistants they actually use, so you're seeing the pattern instead of one lucky or unlucky answer.