How Stacker Built an AI Recommendation Engine Clients Actually Trust

Ken Romano, SVP of Product at Stacker, talks about the AI workflows his team has built, both for internal use and as features in the product.

The first story is an internal assistant for querying data. It flopped at launch because people opened a blank box and didn’t know what to ask. It caught on after the team added a Slack integration and a set of example questions.

Then Romano gets into Sparks, Stacker’s new AI tool that helps brands decide which stories to write. The team built it over 90 days with Claude and a lot of prompt refinement. Sparks looks at a brand’s data and at how stories perform across Stacker’s network, and sends story recommendations every week.

On adoption, he argues for finding internal champions instead of ordering people to use AI from the top. In a business that runs on trust, he says, you have to be deliberate about where AI goes. He prefers starting with controlled, limited uses over open-ended chatbots, and he thinks culture change matters as much as the tooling.

Bring your agent. Bring your team.