AI agents for data & analytics, scored.

Six dimensions, one radar per agent. Every score comes from public documentation and links to its source.

Based on public documentation, not hands-on testing ·

Being scored next: ChatGPT, Claude, Gemini in Sheets, Copilot in Excel, Databricks Genie, Julius AI.

All scores as a table

How we score

Each agent is scored 0 to 5 on six fixed dimensions, always in the same order, with bigger always better. We score the cheapest publicly priced plan that unlocks the agent's data features, and name it on the card. If the documentation doesn't say, the score is n/d (not disclosed) and the radar shows a gap, never a zero.

Scores reflect vendor documentation on the date shown, taken at face value. Hands-on testing with a fixed test pack is coming next and will be labelled "Tested". Nothing here is paid placement.