AMEC, the Association for Measurement and Evaluation of Communication, has launched a set of GEO Principles aimed at bringing rigour to how communicators measure their work in AI-led channels.

The move targets a problem practitioners know intimately. As audiences shift from searching traditional engines to asking chatbots and AI assistants, comms teams face a new question: did our message make it into the machine's answer? The GEO Principles — GEO standing for generative engine optimization — are AMEC's attempt to give the industry a common framework rather than leaving every vendor and agency to improvise its own scoring system.

That improvisation has become a real headache. Over the past two years, a crowded field of monitoring tools has rolled out AI visibility products, each with its own methodology, panel sizes, and definitions of what counts as a "mention" inside a generated answer. Claims vary widely from vendor to vendor for the same brand in the same week. Without shared standards, in-house teams struggle to benchmark, and agencies struggle to prove impact in a way clients can trust.

AMEC is, on paper, the right body to intervene. Its Barcelona Principles have served since 2010 as the industry's baseline for measurement discipline — most notably by insisting practitioners tie communications outcomes to business results rather than count clips. The organisation has periodically updated that framework as channels evolved. The GEO Principles extend that effort to generative AI, where the old proxy metrics break down almost entirely: there is no impression count, no share of voice in any agreed sense, and the same prompt can return a different answer on the next run.

Sceptics will note what this launch is not. It is not a certification, an audit scheme, or an enforcement mechanism. Principles are only as strong as the vendors and agencies that voluntarily adopt them and submit their methodologies to scrutiny. Whether the major AI visibility platforms sign on — and whether they open the black boxes enough for third parties to verify their numbers — will decide if this becomes a genuine standard or another PDF that sits on the shelf next to every other framework.

Still, the timing is not accidental. Client-side demand for AI visibility reporting has exploded, and procurement teams are starting to ask hard questions about what exactly they are buying. A common set of principles gives corporate comms directors something concrete to demand from their agencies and tools: transparent methodology, consistent definitions, repeatable measurement, and honest reporting of the volatility inherent in generative outputs.

For practitioners, the practical next step is straightforward. Read the principles. Put them in front of whoever sells you AI measurement. Ask, specifically, how their product maps to each principle — and keep asking when the answers are vague.

AMEC members will presumably get early guidance on implementation, and the wider test comes over the next year: adoption, methodology disclosure, and whether any vendor's numbers hold up when measured against the framework. The industry has been here before with social media measurement, and it took years to move from vendor-defined metrics to anything approaching comparability. Generative AI measurement now faces the same climb, and AMC has just put a flag at the base of it.