I am launching a short series of videos illustrating what AI agents can do when connected to a well-structured time series platform. Each video focuses on a single use case, showing the platform and an AI assistant working through a real task. The first two are available below.
Connecting an AI agent to the Refinery
The opening video is a two-minute walk-through of the setup. The AI assistant connects to the Timeseries Refinery via MCP, the Model Context Protocol. Once connected, it can read, transform and analyse time series autonomously.
Time series data sits at the centre of most operational decisions — sales trends, energy consumption, financial forecasts. Extracting insight from it has traditionally required a chain of specialists: one person to query, another to clean, another to chart, another to interpret. With the right structure underneath, an AI assistant can compress that chain considerably.
The video is accessible to a non-technical audience. Its purpose is to show what the setup looks like in practice, end to end.
Finding a series in a catalog of hundreds of thousands
The second video addresses a problem of scale. When a catalog contains hundreds of thousands of series, manual search ceases to be a viable option.
The AI assistant relies on the Refinery's metadata and naming structure to surface the right series within seconds, while remaining inside the platform's access rights at every step. Nothing leaves the governed perimeter, and search gains in speed without losing in rigour.
What follows
Further videos will be released in the coming weeks, each focused on a specific business use case. The underlying point will remain the same throughout the series: AI delivers immediate, tangible value when the data layer beneath it is structured, traceable and governed. Without that foundation, the conversation remains theoretical.
Request a demo