Why AI rank tracking is broken, and what to measure instead

Your content team might be having its best quarter ever, and the dashboard will tell you nothing happened. In this episode of Content Amplified, Tony Pataky, Director of SEO at Procore, explains why measuring your brand's visibility inside ChatGPT, Claude, Gemini, and Google AI is fundamentally noisy, and what to track instead. He walks through a SparkToro study from Rand Fishkin that ran roughly 3,000 prompt tests across ChatGPT, Claude, Gemini, and Google AI with about 600 volunteers, and found that less than one in 100 times did ChatGPT return the same list of brands twice, and less than one in 1,000 times the exact same rankings. Tony explains why "rankings" do not really exist in LLMs, why building prompt sets from your high-converting SEO keywords misleads you, and why staying broad and high-level tells you more. He then lays out a three-layer measurement approach: AI visibility tracking, referral traffic, and post-conversion surveys, where some teams find AI was part of 30, 40, or 50 percent of buyer journeys even when LLM referral traffic reads under 1 percent. Listen for a practical way to build a measurement baseline now and reframe the attribution conversation with your CMO.About TonyTony Pataky is the Director of SEO at Procore, where he leads the SEO team along with the marketing performance team covering analytics and CRO. He joined Procore through its acquisition of Levelset, where he worked before the acquisition. His career has been focused on SEO across enterprise companies, startups, and everything in between.Show NotesConnect with Tony on LinkedIn: https://www.linkedin.com/in/tonypataky/SparkToro study by Rand Fishkin on AI answer consistency (roughly 3,000 prompt tests across ChatGPT, Claude, Gemini, and Google AI)Text us what you think about this episode!