Data Story

The Problem

The team set out to build a data analysis agent — something to lower the floor and raise the ceiling for data analytics, letting people explore data in plain language and turn the results into something they could share.

But the project kept stalling on itself. Stakeholders held competing visions for the agent's primary use case, which sent every meeting into the same circular debate. And without a concrete artifact to point at, leadership couldn't see the core user journey or feel the value the team was trying to deliver. The idea was abstract, and abstract ideas don't get funded or built.

The Solution

I made the vision tangible: a concept sizzle reel that turned the abstract debate into one shared picture of the product. It walked through the agent doing real work —

• Explore and ask a simple question in natural language.
• Control the context shared with the agent, so its answers stay grounded.
• Build and share a data report, or turn it straight into a dashboard.

Concretely seeing the journey end-to-end is what broke the stalemate. Once people could watch the agent work, the scope debate resolved itself.

How I got there

• Design sprint to force-rank the competing ideas and narrow scope to the few highest-impact user flows.
• Rapid prototyping — an interactive prototype that demonstrated the agent's key capabilities in a realistic analytics scenario, not a toy demo.
• Storytelling — a narrated video that built excitement and secured leadership buy-in, so the work had air cover to ship.

The Impact

The agent shipped. It launched in Fishfood to hundreds of pilot users and grew into a tool now used by roughly 3,000–6,000 people to compose, share, and store data narratives directly from the dashboards they're built on.

"An extremely nice agent — thanks to everyone who has made this possible!"

"I've been having lots of fun with it. Thank you for this!"

Selected screens available to walk through in conversation (work is under NDA).