Sigma Computing review: Analytics, apps and AI agents built on live data in your cloud warehouse

What Sigma Computing does

Sigma is a platform for building dashboards, applications and AI agents on top of data that remains in your cloud data warehouse. Teams can start from a question in plain language, a spreadsheet-style interface, SQL or Python, and the vendor says IT connects governed data once so other teams can build on it.

Access rules come from the warehouse itself. The homepage states that each person's queries run under that person's own permissions, down to individual rows, and that audit logs record who built, changed or asked what, AI conversations included. Usage dashboards show which workbooks and people create the most load.

The feature list covers natural-language questions, dashboards, formatted reports sent out in batches, and input tables that save spreadsheet-like edits into the warehouse. Apps can trigger actions in Salesforce, Slack or custom APIs. Agents can run on a schedule, watch for changes, write results back and keep a person approving their proposals.

Sigma can also be embedded inside other products through iFrames or the Sigma Rest API, with custom themes and multitenant management. Other entry points include Sigma in Slack and a command-line tool.

The pages offer a free trial and a demo request, but they give no plan names, prices, trial length or usage limits.

• Natural-language questions on live data
• Spreadsheet-style input tables with writeback
• Row-level security from the warehouse
• Scheduled AI agents
• Embedded analytics with custom themes
• Multitenant management
• Audit logs for builds and AI conversations
• Actions to Salesforce, Slack or custom APIs
• Batch-delivered formatted reports
• Sigma CLI

Our take

Sigma looks suited to organizations that already keep their data in a cloud warehouse and want business teams to build dashboards, writeback apps and agents without copying data elsewhere, while IT keeps permissions and audit trails. The embedding and multitenant features also address software companies that want analytics inside their own product. It is a weaker fit if you need to see prices before talking to anyone, because no plans, prices or trial terms appear on the official pages. It also assumes a warehouse is already in place, so it does not read as a standalone tool. This view rests only on the vendor's own pages.