I've been running GEO, measuring and shifting how AI engines talk about brands, for close to two years. Known & Cited, the consultancy built on that work, is six months old. This page is the honest account of both timelines: what the practice taught me before the company existed, and what the company has learned since.
The work started before it had a name. In 2024 I was watching AI answers replace search results for the brands I worked with, and nobody could tell me what AI was actually saying about them. So I started measuring it by hand: asking the engines the questions real buyers ask, logging the answers, tracking who got named and who got cited. It was crude and it was revelatory. Brands with strong SEO were invisible to AI. Brands nobody had heard of were being recommended because one comparison site loved them. The rules were different, and almost nobody was reading them.
Known & Cited exists because that measurement gap never closed on its own. By early 2026 there were plenty of tools that would give you a dashboard, and almost nobody who would give you a defensible answer: here is where you stand, here is why, here is what to do, in that order. A general PR or marketing consultancy can't do it, because the measurement layer is real technical work. A software tool can't do it, because the judgement layer is real consulting work. K&C is built to do both.
Three things. We split the offer into three products, AI See, AI Think and AI Do, because one score across the whole funnel was hiding more than it showed. We adopted the AMEC GEO Principles as the design spine for every measurement, because the industry finally produced a framework worth aligning to. And we moved to multi-engine measurement as standard, because buyers don't live on one engine and your score shouldn't either.
The methodology holds. Measured over a seven-day window, across around 6,000 responses, AI answers are stable enough to score and track honestly. Source strategy works: when you change what the most-cited domains say about you, the answers follow. And the funnel split is right; a brand can be Cited at the top of the funnel and Overlooked at the bottom, and knowing that changes what you do next.
How fast the engines change, and which changes matter. What genuinely moves Google AI Mode versus what moves Perplexity. How to price ongoing measurement so small brands aren't shut out. We publish what we learn as we learn it, and we say so when we're not sure. GEO is young; anyone claiming certainty is selling it.
More published research under our own name. Our own score, in public, from the end of July 2026. And a steady widening of the sector estate as more categories start asking the questions we measure. The company is six months old. The practice is two years old. Both are just getting started.
Written by Russ Read-Barrow, founder of Known & Cited. Bath, July 2026.