I was recently reviewing a product flow when something clicked. The process had seventeen distinct steps, each asking users to make decisions about features or services they hadn’t experienced yet. Our completion rate was dismal.

I recently sat in a strategy session where a team spent hours debating whether to rebuild their underlying backend engine using the “latest and greatest” framework. The discussion covered algorithms, infrastructure costs, and competitive benchmarks.

Every week, another vendor slides into your DMs with promises of “revolutionary agentic AI solutions” that will “transform your entire business.” The marketing drumbeat is deafening: Gartner names agentic AI as the top strategic technology trend for 2025, IBM declares 2025 “the year of the agent”, and McKinsey positions agents as “the next frontier” of AI innovation.

Ever notice how the most powerful question in product management is also the favorite of three-year-olds, philosophy professors, and 90s boy bands?

McKinsey’s recent article “The missing data link: Five practical lessons to scale your data products” has hit on something I’ve been advocating for years: treating data as a product rather than a project or resource. While I’m typically wary of consultancy research (aren’t we all?), this piece validates much of what we’ve observed in the field about the transformative power of data product thinking.

We’ve all seen it—-companies rushing to slap “AI-powered” on their marketing materials while implementing little more than basic automation or, worse, glorified if/then statements. The current marketplace is brimming with products that tout AI capabilities but deliver minimal value beyond the initial novelty.