Data analysis agents
Agents that query, clean and explain data, and produce reports and charts.
15 agents, each with its inputs, outputs and review level checked by a person.
Data analysis agents answer questions about your data in plain language: they write and run the queries, produce charts and tables, explain what the numbers show, and in some cases build reports on a schedule. They connect to a warehouse, a database, a spreadsheet or an uploaded file. The listings record what each agent connects to, what it returns, and whether an analyst reviews the result before it reaches a decision maker.
Choose by where your data lives and who will ask the questions. Agents built into a warehouse or a BI tool inherit its access controls and metrics definitions, which matters for consistent answers. Standalone notebooks and spreadsheet agents suit ad hoc analysis on files. In every case, check what the agent does when a question is ambiguous or the data is incomplete; the better contracts keep unknowns unknown.
The agents
- Akkioby AkkioMedia analytics agents that segment audiences, forecast campaigns and report on performance.Quote on requestNeeds configurationOptional reviewDetails
- Copilot for Power BIby MicrosoftChat with Power BI reports and semantic models to get summaries, answers and draft report pages.Usage-basedNeeds configurationHuman reviewDetails
- Databricks AI/BI Genieby DatabricksAnswers business questions in plain language from data governed in Unity Catalog.Usage-basedNeeds configurationOptional reviewDetails
- DataLabby DataCampAI data notebook that writes SQL and Python from a question and returns a shareable report.FreemiumNeeds configurationHuman reviewDetails
- Deepnote AIby DeepnoteNotebook agent that plans, runs SQL and Python, and deploys the result as an app or API.Freemium, from $39 per editor/mo billed yearlyNeeds configurationHuman reviewDetails
- Defogby DefogAnswers questions about structured databases by generating SQL with the SQLCoder model.Subscription, from $5,000/moDeveloper setupOptional reviewDetails
- Hex Notebook Agentby HexWrites and debugs SQL and Python in a Hex notebook from a plain-language question.Freemium, from $36 per editor/moNeeds configurationHuman reviewDetails
- Lumi AIby Lumi AIChat-based analyst that builds a semantic model from enterprise data and surfaces insights on its own.Subscription, from $50/moNeeds configurationOptional reviewDetails
- Rows AI Analystby RowsSpreadsheet AI that writes formulas, builds charts and analyses data from a typed prompt.Freemium, from $8 per user/moWorks instantlyHuman reviewDetails
- Snowflake Cortex Analystby SnowflakeTurns a business question into SQL against Snowflake data using a semantic model.Usage-basedDeveloper setupOptional reviewDetails
- ThoughtSpot Spotterby ThoughtSpotAnalytics agent that answers business questions with traceable queries and can trigger actions.Subscription, from $50 per user/mo billed annuallyNeeds configurationOptional reviewDetails
- Vanna AIby Vanna AIOpen-source SQL agent that answers questions about company databases in multi-turn chat.Subscription, from $50/moDeveloper setupOptional reviewDetails
- Vizlyby VizlyUploads a data file, answers questions in plain language and writes the Python or R behind each chart.FreemiumWorks instantlyHuman reviewDetails
- Wren AIby Wren AIAnswers business questions from over 20 data sources through a governed semantic context layer.Freemium, from $179/mo billed annuallyNeeds configurationOptional reviewDetails
- Zenlyticby ZenlyticAnalytics agent that investigates a data question and delivers cited decks, memos and models.Pricing not publishedNeeds configurationOptional reviewDetails
No agent matches. Clear the search or choose another filter.
Questions
Do I need to know SQL to use these agents?
No. The point of the category is asking in plain language. Several agents show the generated query so a person can check it, which the listing notes where the vendor documents it.
Are the numbers trustworthy?
As trustworthy as the query and the data. Agents that show their work and respect defined metrics make errors easier to catch. The review level on each listing shows whether a person checks results before they are shared.
Where does my data go?
That depends on the vendor and is not something this catalogue verifies. Read the vendor's security and data pages linked from the listing before connecting anything sensitive.