Repositioning a legacy service for the Agentic AI Era

The challenge

Statista had real ambition to commercialize its data API, but the foundation underneath it (documentation, developer experience, internal tooling and processes) hadn't yet caught up to that ambition. At the same time, the market research industry was facing a structural shift: as AI makes "results" available a prompt away, the value of traditional data and research products is being redefined, and buyer attention is harder to earn. Statista needed more than a set of recommendations from an outside advisor. It needed a partner who could operate inside the business as a technical entrepreneur, spotting the right emerging opportunities and building toward them quickly.

What we did

ctrl+eval joined as an embedded, interim product lead, acting as an embedded entrepreneur within what has since grown into Statista's "Connect" unit. The engagement spanned modern developer documentation and internal sales tooling; product vision and prioritization work alongside engineering to improve API usability; hands-on discovery, solution engineering, and demos directly with customers, including major enterprise accounts; and early, deliberate bets on emerging distribution channels, most notably a Model Context Protocol (MCP) server and a data partnership with Perplexity, well ahead of the wider market's shift toward agentic AI. In parallel, ctrl+eval pushed evaluation ("evals") infrastructure as a core product direction and helped grow the team behind it.

Approach

The work moved in deliberate phases rather than a single fixed roadmap. It began with discovery and the basics: modern documentation and Retool-based tooling for the sales team, then moved into sharpening product vision and working directly with engineering to improve API usability and identify priority use cases. From there, the focus shifted to prototyping and championing MCP as a product direction, growing the engineering team, and helping shape what became the Connect department. Later phases built out ETL and connectivity infrastructure and helped land the Perplexity data partnership, with ctrl+eval supporting the commercial and strategic side. Followed by pushing an evals suite as a new product direction and stepping into an solutions engineering role on major enterprise accounts. Throughout, the approach stayed hands-on and close to customers rather than advisory from a distance.

Working with Adam was one of the most energizing experiences I've had bringing in external expertise. As our interim product lead, he hit the ground running, arriving with a sharp read on where the market was heading and an almost uncanny ability to filter out the noise and zero in on what actually mattered...

Learnings

Obsess over customer experience and empathy

The clearest signal for where to build next came from staying close to customers, through discovery calls, demos, and solution engineering, rather than from internal roadmaps alone.

Building is cheap in 2026: prototype and experiment more

Rather than over-planning, the team prioritized getting working prototypes (like the MCP server) in front of the market early, using real signal to decide what to scale.

AI is redefining what market research buyers pay for

When answers are a prompt away, incumbents can't rely on legacy distribution. Betting early on AI-native integration points (MCP, agentic partnerships) was less a technical experiment than a strategic repositioning of the business.