AMEC has launched a set of GEO Principles to bring measurement rigour to AI-led discovery, after a six-month development process involving senior practitioners from PR Agency One, Ketchum, FleishmanHillard, Hotwire Global, Big Valley Marketing and Converseon.

The move responds to a measurement problem every practitioner now faces: AI-generated summaries, conversational search and zero-click discovery increasingly decide how organisations, brands and issues are found, understood and trusted online. GEO — Generative Engine Optimisation — has become shorthand for how organisations show up in AI-generated answers.

AMEC's stated aim is to stop measurement collapsing into simplistic rankings, vanity metrics or opaque scores from whatever tool a vendor happens to be selling this quarter. Given how much GEO tooling has appeared with little methodological transparency, that caution is warranted.

What does the framework actually measure?

The principles set out a practical framework covering three connected areas:

• Upstream reputation signals — earned coverage, third-party commentary, reviews, expert content and owned assets;
• Search and content readiness — whether an organisation's digital presence is credible, accessible and structured for interpretation by search engines and AI systems;
• Downstream AI outputs — how an organisation appears in AI-generated answers, citations, framing, omissions and potential reputational risk.

The principles also introduce baseline evidence requirements: repeatable prompts, documented methods, transparent assumptions and clear limitations. AMEC reinforces that AI outputs should be treated as directional evidence rather than absolute truth, and warns against relying on any single score, platform or tool.

Who built it?

James Crawford of PR Agency One led development alongside Mary Elizabeth Germaine of Ketchum, Ben Levine of FleishmanHillard TRUE Global Intelligence, Matt Oakley of Hotwire Global, Amber Daugherty of Big Valley Marketing and Rob Key of Converseon.

Crawford did not mince words about the current state of the market.

"Anyone working in PR or communication will know how quickly clients and boards have started asking how GEO and LLM outputs should be measured. There is excellent innovation taking place, but there are also uneven standards, overclaiming, vanity metrics and methodologies that are not always transparent enough," Crawford said.

"These principles give the industry a more rigorous way of looking at AI-led discovery: one that recognises its importance, but also its limits. The most useful measurement will come from triangulating evidence: the reputation signals that feed the information environment, whether organisations are technically and editorially discoverable, and what AI systems then present to users."

That triangulation argument is the framework's core claim — and the one practitioners will need to test against client pressure for a single number.

Why now?

AMEC CEO and Global Managing Director Johna Burke framed the initiative as an accountability issue for the whole industry.

"As AI increasingly shapes what people see, trust and act upon, the communication industry must hold itself to higher levels of transparency, evidence and accountability," Burke said.

"This initiative reflects the collective expertise, scrutiny and commitment of professionals across regions who understand that rigorous, transparent and ethical evaluation is essential to maintaining trust in the AI era."

For measurement teams, the immediate practical takeaway is procedural: document your prompts, publish your assumptions, state your limitations, and never let a single tool's score stand in for evidence. Whether agencies and vendors adopt those baseline requirements — or keep selling black-box GEO dashboards — will determine whether the principles have teeth.