How we measure: the K&C methodology

Known & Cited measures how AI engines cite and recommend brands using a repeatable, defensible methodology built on the AMEC GEO Principles. Around 6,000 AI responses per report, four engines as standard, five bands per product, and a K&C analyst in the loop at six named checkpoints. This page explains the whole thing in plain English.

Apply for a free Discovery Exec Brief → See the worked examples

The AMEC GEO Principles

Every K&C measurement is designed around four dimensions: personas (who is asking), journey stages (where they are in the buying process), sub-products (which part of your offer they are asking about) and buying factors (what actually drives their choice). Multiplied together, these produce the question set your real buyers put to AI. We measure the answers to those questions, not a generic prompt list.

The three products

The funnel splits into three measurable questions, and K&C has a product for each. AI See: am I named at all? AI Think: how am I described? AI Do: am I in the shortlist? Each product carries its own five-band scale, so a score always comes with a plain-English reading of what it means.

Five bands per product

Ghost0-10Whisper11-30Emerging31-50Cited51-75Known and Cited76-100
At Risk0-10Exposed11-30Mixed31-50Trusted51-75Authoritative76-100
Overlooked0-10Noticed11-30Shortlisted31-50Recommended51-75In the Basket76-100

Top to bottom: AI See (red), AI Think (amber), AI Do (green).

Multi-engine measurement, at the tier your buyers actually see

We measure at the consumer scraping tier: the versions of the engines your buyers actually use, not developer APIs that can answer differently. Our standard set is ChatGPT, Google AI Overviews, Google AI Mode and Perplexity. Additional engines, including Claude, Copilot and Grok, are added on client request.

Sampling design

Each full report runs on around 6,000 AI responses by default, sampled over a seven-day window. AI answers change day to day; a single-day snapshot is a coin toss, a seven-day sample is a measurement.

Scoring and weights

The headline score weights three things: AI Visibility at 40%, Source Quality at 30% and Narrative Fit at 30%. The weights are published because a score you can't interrogate is a score you shouldn't trust.

Where humans are in the loop

A K&C report is AI data crunching plus team expertise, not an automated data dump. A K&C analyst reviews, interprets or overrides the machine at six named checkpoints:

  1. Prompt design. The question set is built by hand against the AMEC GEO Principles, then reviewed before anything runs.
  2. Source triage. Every cited domain is classified and sanity-checked by a person.
  3. Sentiment read. Machine sentiment is reviewed against human judgement before it reaches a score.
  4. Recommendation shortlist. Recommendations are ranked by effort versus impact and tagged with an owner, by us, not by a model.
  5. Executive commentary. The narrative sections are written by the team, in our voice, with our judgement.
  6. Final QA. Nothing ships without a full human pass.

Our specialist measurement partner

The measurement layer runs on our specialist measurement partner's enterprise platform, which K&C configures, operates and quality-controls. The methodology, the scoring, the bands and the recommendations are K&C's own.

The honesty box

GEO is a young discipline. We do not guarantee outcomes, we do not promise rankings, and we will tell you when the data is too thin to call. What we guarantee is a defensible measurement, a clear reading of it, and recommendations we would stake our own name on. We are honest about what we know and what we don't.

"Our measurement is designed so that a sceptical CFO can pull any thread and find method, not magic."
Russ Read-Barrow, founder, Known & Cited, July 2026

Related pages

%s