Skip to content

AMEC launches AI measurement principles amid rise of GEO

1 min

As AI-generated answers reshape online discovery, communications groups are racing to measure how brands appear in large language models.

By Rebecca Pardon

  • Data & Insights

The International Association for the Measurement and Evaluation of Communication (AMEC) has launched a new framework intended to bring greater transparency to the emerging market for measuring AI-generated search and discovery results.

The AMEC GEO Principles, unveiled at the AMEC Summit on Tuesday, aim to help communications professionals assess how organisations are represented in outputs generated by large language models and conversational search tools. GEO, or generative engine optimisation, refers to efforts to improve how brands and organisations appear in AI-led search environments.

The framework focuses on reputation signals that shape how organisations are understood online, the accessibility of digital content to AI systems and the analysis of AI-generated outputs, including accuracy and prominence.

James Crawford, managing director at PR Agency One and an AMEC board director, led the initiative. He said the rapid growth of GEO had created both innovation and scepticism across the industry.

“GEO is moving so quickly and there is both innovation and risks being taken but also rightly a lot of scrutiny and cynicism,” he said. “The risk is that the industry rushes to create a new generation of vanity metrics or black box visibility scores without proper standards, transparency or context.”

Crawford said the principles were intended to encourage a more evidence-led approach to assessing AI-driven discovery, arguing that organisations should not focus solely on whether a brand appears in an AI-generated answer.

“Good GEO measurement therefore cannot simply be about whether a brand appears in an AI answer as measuring that is unreliable,” he said. “It has to examine why it appears, how accurately it is represented and what sources and narratives shaped that output.”

The principles were developed over six months with input from agencies, academics and technology providers, including contributions from Ketchum, FleishmanHillard and Hotwire Global.