Agentcy opens free AI visibility module for accuracy

Agentcy has open-sourced the AI visibility module of its communications platform and made it free to use. The software is intended to help agencies and communications teams assess how brands are represented across major AI models.

The release makes a measurement tool publicly available at a time of growing scrutiny of search and AI ranking systems. It is designed to show not only whether a brand appears in AI-generated answers, but also whether the description is factually correct when checked against approved source material.

The London-based company, which sells software for public relations teams, is targeting a growing concern for agencies and corporate communications departments. As AI search tools and answer engines become a more prominent route to information, brands face a different challenge from traditional search rankings: they may appear in responses while still being described inaccurately.

According to Agentcy, the software breaks brand descriptions into individual claims and checks them against a verified source of truth. It then flags differences such as incorrect facts, outdated positioning, missing context and confusion with competitors, while linking each score to the original prompt, answer, cited web sources and timestamp.

The module also tracks how often a brand is mentioned or recommended across leading AI models, whether cited sources lead to an actual brand mention, the tone of responses, how competitors are presented in the same answers and how those patterns shift over time.

Accuracy focus

Agentcy is trying to distinguish its approach from a crowded market of AI visibility and search optimisation tools that have largely focused on mentions, citations and sentiment. It argues that those measures do not show whether an AI system has got the underlying facts right.

The argument comes as marketers and communications specialists work out how to judge brand presence in AI-generated answers. In conventional search, position on a results page has long been a core metric. In AI search, where a model can summarise, omit or distort information, agencies are increasingly looking for ways to test not just prominence but correctness.

Stuart Bruce, an industry commentator and adviser on AI communications, said the distinction matters for reputational work. "Much of the hype about GEO centres on visibility but for the work we do with clients in the corporate, government and not-for-profit space then being portrayed accurately in AI answers is even more important. Tools that just focus on visibility can't provide corporate communications professionals with the data they need to protect and improve reputations," said Stuart Bruce, the PR Futurist and co-CEO of AI communications transformation consultancy Purposeful Relations.

Industry pressure

Agentcy's move also reflects pressure on vendors to show how their tools produce results. It cited market concerns over expensive AI visibility products, inconsistent findings and limited evidence that such products can prevent hallucinations or misattributions in generated answers.

By publishing the measurement layer under the MIT licence, Agentcy is giving agencies a tool they can inspect and adapt themselves rather than relying on a proprietary scoring system. The software can run locally or on an agency's own infrastructure, according to the company.

That open approach may appeal to public relations firms and in-house teams that need to explain their methodology to clients, senior management or regulators. In communications work involving listed companies, government bodies or charities, the ability to show how a conclusion was reached can carry as much weight as the score itself.

The wider debate has gained momentum as regulators look more closely at the systems that determine online visibility. Search ranking transparency has already become a competition issue, and similar questions are emerging around AI-generated answers, including how models choose sources, frame narratives and what recourse organisations have when they are portrayed inaccurately.

Open method

For agencies, the practical issue is whether they can build a repeatable method for checking AI answers across different models. Agentcy said the module was developed with input from senior practitioners in public relations, corporate communications and public affairs, with the aim of offering a shared and auditable framework.

Tom Fry, the company's chief technology officer, said the market had focused too narrowly on whether brands are present in AI answers. "Visibility tells you whether you're in the room. Accuracy tells you whether what's being said is true."

He argued that the economics of current tools are pushing agencies towards methods they cannot fully validate. "The market is telling us something," Fry said. "Agencies are paying for tools that measure whether a brand appears in a result. Not whether what it says is correct. That is the gap we built this to address. Flying blind is expensive. Open-sourcing the measurement layer means any agency can verify the results themselves, without trusting a black box they did not build and cannot inspect. If PR is going to be taken seriously as a performance discipline, accuracy has to be part of what we measure."