The four stages, in order
Each stage answers a question the next one needs. Skip the first and you’re guessing at the problem. Skip the last and you’ll never know whether the middle two did anything.
We find out which engines your buyers actually use, and what each one says about your category right now.
- We agree the questions with you first. The ones someone types when they’re choosing a supplier, not reading around the subject.
- Every question goes to all six engines, signed out, more than once. One run tells you almost nothing.
- Every answer gets marked the same way: named with a source, named without one, or not there at all.
Rules out The idea that checking ChatGPT tells you where you stand. The six disagree constantly. One averaged score won’t tell you which is costing you business.
CWe work out why you’re missing. There are three possible reasons and they need completely different fixes.
- It can’t identify you. The AI is unsure who you are or what you sell, so it won’t put your name forward.
- It can’t find anything about you. It knows who you are, but none of the sources it trusts mention you.
- It prefers someone else. It knows you, the mentions exist, and it still names two competitors. This is the hard one, and we’d rather tell you early.
Rules out The assumption that one fix covers it. Publishing more pages helps with the second problem and does nothing for the other two. That’s why “just write more content” is the advice everyone gives and nobody gets much from.
AWe do the work, and which work we do depends entirely on what stage two found.
- We make the basic facts about your business consistent everywhere an AI can find them. It stops hesitating over who you are.
- We write pages an AI can quote from directly, instead of pages it has to rewrite and therefore skips.
- We go after mentions on the sites your category’s answers are actually built from. This is the slow part, and the part that moves it.
Rules out The assumption that this is a website job. Most of an AI answer comes from what other sites say about you, so most of the work happens somewhere you don’t control.
NEvery month we show you what moved because of us, with your category’s own movement taken out of the figure.
- We set aside a group of your questions at the start and never touch them. That gives your category somewhere to show its own movement.
- Both groups get measured again together. Same day, same engines, same conditions.
- The gap between them is your net gain, and it’s the first number in the report.
Rules out A number that went up being sold to you as a result. These engines move on their own. Models get updated, competitors do things, sources get re-indexed. Part of any rise was never down to your agency. Everyone else reports the rise. We report the share we caused.