Data-driven insights: when data becomes the product

By Rogerio Melfi, co-founder and CPO of PilotIn.

The promise of turning data into value has never been more concrete. And this is not just rhetoric: in 2025, the global volume of indicators created and moved is expected to exceed 180 zettabytes, according to Statista — equivalent to an average of more than 340 terabytes generated per minute. However, the way companies view this transformation still varies drastically. Some treat artificial intelligence as a cost, others as an efficiency tool, and a few — still very few — see it as a direct source of revenue.

This journey is not merely technological; it is strategic. In the past, innovations were used to automate operational tasks, save time and trim teams. It was a response to pressure to cut expenses. Then we evolved to a second layer: AI and data as optimization accelerators. Products began predicting behavior, automating journeys and reducing churn. We started measuring what was previously invisible. And with that came impact: according to McKinsey, data-driven companies — organizations that strategically use codes to guide their decisions — are 23 times more likely to acquire new customers and 19 times more likely to be profitable.

But now we are entering a new era, in which intelligence starts generating profit and returns directly, rather than merely saving resources. This leap became possible only because the computing foundation changed. The old Moore’s Law, which for decades predicted a doubling of transistors every two years, may have slowed physically, but it has been replaced by exponential leaps in the power of chips, cloud clusters — groups of computers working together — and architectures optimized for AI. Model training costs have fallen, and response latency has become almost imperceptible to end users.

With this foundation, artificial intelligence and data no longer merely support the product — they are the solution itself. When a personalized recommendation becomes a conversion. Or when the categorization engine becomes a competitive advantage. When predictable financial behavior leads to better credit decisions, and that becomes revenue, not just efficiency. At this point, intelligence stops being “support for the development team” and becomes the core of the value proposition.

Making this happen requires more than dashboards and well-tracked events. It requires data design at the source, from the very first click. A culture that connects this area to decisions. Hypotheses that become experiments. Teams able to move from observation to action. Indicator-driven products do not emerge from reports — they are born from living architecture, carefully maintained pipelines, traceability and an obsession with relevance.

Moreover, they are born from a clear vision of impact. In a landscape where 81% of product leaders report that indicator-based decisions outperform intuition-driven ones, according to Product Collective, insisting on “gut feeling” means falling behind.

Ultimately, every digital product will encounter these three stages: AI as cost, as efficiency and as revenue. The question is: where is your company today? More importantly, is it building the right data to reach the next stage?

📌 Originally published in Startups