New AI audit helps executives manage how chatbots describe them

The risk of AI first impressions

When business leaders prepare for a fundraise, a board appointment, or a media interview, they often rely on AI assistants to do the preliminary research. These tools, including ChatGPT and Gemini, are increasingly used to summarize who a person is and what they have achieved in their careers. However, this new method of discovery introduces a significant risk for high-level executives.

The problem lies in the data sources these models use to build their answers. They may pull from an outdated biography or, even worse, provide information about a completely different person with the same name. As The Daily Tribune News reported, this means a leader's first impression with a potential partner might be based on incorrect or irrelevant information.

Ben Yeoh, the founder of ProfileArchitect, noted that most executives have never actually checked what these AI models say about them. He argues that because AI now writes the first impression, leaders need a systematic way to measure and maintain their digital presence. His new service, the Signature audit, is designed to reveal these hidden inaccuracies.

Testing the AI footprint

To demonstrate the problem, Yeoh used himself as a test case for a 30-question study involving ChatGPT, Gemini, and Perplexity. The results showed that all three models provided at least one answer about him that actually belonged to someone else. For example, Gemini identified him as a Malaysian automotive chairman, while Perplexity provided information entirely about a London-based investor.

The Signature audit works by running 20 to 30 specific prompts across these major platforms to see what they output. It also performs an identity check across various knowledge panels, Wikidata, LinkedIn, and published biographies. This helps identify where the AI is getting its facts and where it is getting confused by namesake profiles.

The audit concludes with a four-layer scorecard that evaluates owned sources, name distinctiveness, current AI outputs, and third-party authority. It provides a report that identifies the highest-leverage fixes for the executive. This process is intended to move beyond simple reputation management by creating an authoritative source that AI assistants can trust and cite.

Who is affected by AI search

This service is specifically designed for CEOs, founders, and board directors who rely on their names to open doors. For these individuals, a wrong description isn't just a minor error; it can be a barrier to significant business deals or professional opportunities. The audit aims to ensure that the digital architecture surrounding a leader is accurate and consistent.

Yeoh brings a background in corporate training and performance coaching to this work, having worked with over 7,000 people. His experience with large organizations like DHL, Petronas, and Unilever informs his approach to reputation. He views the management of a digital footprint as a system that requires regular measurement and maintenance rather than a one-time fix.

The audit is currently open by application for executives across the Asia-Pacific region. While the service is a paid engagement, it is positioned as a small cost compared to the potential damage of a wrong first impression. It offers a structured way for leaders to take control of their narrative in an era where AI is the primary gatekeeper of information.

The roadmap for accuracy

The Signature audit is priced at SGD 1,500 and takes approximately 10 business days to complete. Once finished, it provides a 90-day roadmap to implement the necessary changes to the executive's online presence. This roadmap helps leaders prioritize which sources to update first to ensure the AI models eventually reflect the correct information.

ProfileArchitect also offers an owned executive profile build called Vantage, which the audit credit can be applied toward. The goal of this work is to build 'owned' profiles rather than 'rented' ones. By creating a dedicated, AI-readable site, leaders can provide a primary source of truth that search engines and AI tools can easily find and verify.

Yeoh emphasizes that no provider can guarantee exactly what an AI will say at any given moment. Instead, his method focuses on recording what the AI says today, fixing the underlying sources it relies on, and then re-running the tests. This creates a visible record of progress, allowing leaders to see their digital footprint improve over time as the AI updates its knowledge.