We talk with Matt Neligeorge, Bobsled's Chief Scientist, about why agentic analytics is hard, why a good harness beats a better model, and how background agents that learn from usage are improving accuracy on the questions that matter most.

As agent models keep improving — moving from tools a human wields to systems that run long, multi-step tasks on their own — we're heading toward agents that curate their own context, learning from every question asked, every query run, and every piece of feedback received.

Three product launches — the Analytics MCP, Data Skills, and Workflows — plus a stack of quality-of-life upgrades.

The flexibility of an agent. The predictability of a pipeline.

Repeatable prompts that let data teams teach agents how to handle complex analyses on demand — codified once, reusable forever.

Most agentic analytics POCs survive the demo. Few survive context drift, and the problem only gets worse when agents, not people, are the ones asking the questions.

A governed endpoint that gives agents an analyst on the other side of the connection, not just an API. Built for data teams that want to support agentic workflows without losing control of how their data is interpreted.

We talk with Ben Tengelsen, Director of Data Products at Aampe, about why data sharing is core to Aampe's platform, how the team thought about build versus buy, and where analytics is headed next.