Tony Uphoff Tests AI Search Engines To Identify Measurement Gap

Tony Uphoff, a five-time CEO, conducted an experiment to test how AI search engines recommend businesses, according to Search Engine Journal. He asked ChatGPT, Gemini, Perplexity, and Google AI Overviews questions that prospective clients might ask, and then counted how often the engines recommended his own businesses.

Uphoff's test showed that while the AI models knew who he was, they rarely recommended his businesses, highlighting a measurement problem in the industry. This gap between being known and being recommended may be a concern for many brands, as it can affect their visibility and valuation.

Search Engine Journal reported that Uphoff's test results were humbling, with his newsletter and firm appearing in only a few of the answers. However, when asked about his career, the engines provided accurate information. This discrepancy suggests that being known is not the same as being recommended.

Uphoff's argument is that the way businesses measure their online presence is flawed, as it focuses on discoverability rather than consideration. He notes that for two decades, the goal of SEO has been to rank high in search results and earn clicks, but this approach may not be effective in the age of AI search.

The survey data supports Uphoff's argument, with 51% of B2B software buyers saying they now start their research with an AI chatbot, and 69% saying a chatbot led them to choose a different vendor. This shift in behavior may require businesses to rethink their online strategy and focus on being recommended by AI models.

Another finding from Uphoff's test is that answer engines often default to outdated information about a company, which can be costly. This can happen when a company has undergone a rebrand, merger, or leadership change, and the old information is still available online. Uphoff notes that few teams are actively tracking and measuring AI visibility, which can lead to a stale record of the company.

Uphoff's test also showed that AI shortlists are often built from other people's shortlists, rather than from original research. This means that companies may not be credited for their own ideas and innovations, but rather for publishing clear explanations of them. This can have implications for content strategists and marketers.

Finally, Uphoff's work highlights the importance of owning customer relationships, rather than renting them from search and platforms. He notes that buyers are willing to pay more for businesses that have a direct relationship with their customers, and that AI disruption protection is essential for business valuation.

Overall, Uphoff's test and arguments suggest that businesses need to rethink their approach to online presence and measurement in the age of AI search. By focusing on being recommended and owning customer relationships, businesses can improve their valuation and stay ahead of the competition.