The analytics interface is changing

For a long time, self-service analytics meant giving more people access to dashboards and, in some cases, governed tables so they could query the data themselves.

LLMs are starting to change that model.

Self-service today can increasingly come in the form of natural language, with organisations expecting data teams to make that capability available across the company.

But that comes with a different set of challenges. It’s not just about plugging an LLM into your warehouse. It needs business context.

Anthropic’s self-service analytics setup with Claude is an interesting example. Users can ask data questions directly in Slack, but the interesting part isn’t really the chat interface. It’s what sits underneath it.

If someone asks why sign-ups have dropped recently, the reason might not exist in a table. It could be in an incident report, a Slack discussion, release notes or a product decision.

The warehouse tells you what happened. Business context helps explain why.

That makes me wonder where self-service analytics is heading.

Maybe it becomes less about giving everyone better dashboards and more about giving people a trustworthy conversational layer across both data and business context.

If that happens, the job of data teams may shift further away from answering every question themselves and towards building the foundations that allow those questions to be answered reliably.