3 Ways Your Agentic Product Is Losing Customers

Arpy Dragffy is the founder and CEO of PH1 Research, an AI strategy consultancy that helps organisations measure and grow the impact and adoption of their AI products. For more than 20 years he has worked across technology, media, sport, finance, healthcare, telecom, retail and the public sector, helping teams prove whether what they built is actually working and fix it when it isn’t. He has led AI adoption and trust research for Microsoft, research for Spotify, and helped launch emerging-tech products for the National Football League and UFC. Executives bring him in to facilitate the decisions about how to put emerging technologies to work, and he hosts the Product Impact podcast.

Most enterprise AI investments underperform for the same reason: the diagnosis stops at strategy and never reaches the work itself.

I study what actually happens at the workflow level – the specific moment where a team decides to use a new tool, or does it the old way. That decision point is where value stalls or compounds. Most organizations never see it, because they're not looking at the right layer.

That layer is under-researched, under-invested, and quietly responsible for most of the ROI failure in AI right now.

For 20 years I've been mapping that gap across products, customer experiences, and AI investments – at organizations like Spotify, Microsoft, Mozilla, the NFL, and the Government of Canada.