Sigma is a warehouse-native BI and AI app platform with agents, writeback, and spreadsheet UX. Free trial offered; pricing via sales. Full review.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
Sigma, from Sigma Computing, is a business intelligence and analytics platform that runs directly on a company's cloud data warehouse. Its website calls it the AI runtime for business: a place to build analytics, dashboards, apps, and agents on live warehouse data, with governance inherited at the source. The company says more than 2,000 enterprises use it.
Rather than copying data into a separate engine, Sigma queries the warehouse live and applies the permissions already defined there. Users can work through chat, a spreadsheet interface, SQL, or Python, and the results land in the same governed workbook. Sigma pitches itself as next-generation BI combined with applications that can write data back.
Sigma connects to a cloud data warehouse such as Snowflake, Databricks, ClickHouse, or a warehouse on AWS, Azure, or Google Cloud. There is no data movement or extract step. Queries run live, and Sigma reuses cached results instead of querying the warehouse again, while usage dashboards show which workbooks and people drive the load.
Security follows each user. Every query runs with the viewer's own warehouse permissions, down to the row, so dashboards, AI answers, and agents all respect the same access rules. IT sets those rules once, and everything built in Sigma inherits them. Audit logs record who built, changed, and asked what, including every AI conversation.
Sigma supports several working styles. Ask Sigma answers natural language questions against the semantic layer and shows how each result was produced. The spreadsheet interface offers 400+ familiar functions, drag, filter, and pivot. Analysts get a full SQL editor and Python notebooks inside workbooks, while business users can drill down and filter with point-and-click exploration.
AI features include LLM functions usable in spreadsheet cells: prompt, classify, sentiment, summarize, translate, JSON extract, web search, and image. Data models define metrics once for reuse. Input tables save edits back to the warehouse. Sigma Agents run on schedules, catch overnight changes, and take approved actions. Embedded analytics and pixel-perfect reports cover customer-facing and operational reporting.
Sigma is a cloud analytics and AI app platform from Sigma Computing. It runs directly on a company's cloud data warehouse and lets teams build dashboards, reports, apps, and agents on live data. Users can work through chat, a spreadsheet, SQL, or Python.
The pages reviewed do not publish Sigma's prices or plan tiers. Buyers can start a free trial, request a demo, or contact the Sigma team to learn about pricing.
Yes. The Sigma website offers a free trial for building analytics and business applications, along with a request-a-demo option.
Sigma connects to cloud warehouses including Snowflake, Databricks, ClickHouse, AWS, Azure, and Google Cloud, and lists AlloyDB as a partner integration. It also works with Slack, Claude, Codex, and ChatGPT, and supports SSO through providers such as Okta and Azure AD.
Sigma says every query runs with the viewer's own warehouse permissions, down to the row, and that audit logs record who built, changed, and asked what. The site lists SOC 2 Type II, PCI DSS, HIPAA, and GDPR, plus SSO and SCIM.
No. Sigma offers a spreadsheet interface with 400+ functions, natural language questions, and point-and-click exploration. SQL and Python are available for analysts and data engineers who prefer code.
Yes. Input tables and writeback let users save edits to the warehouse, and agents can write results back with the user's approval and existing permissions.
Sigma fits organizations that store data in a cloud warehouse and want governed self-service analytics and AI apps across teams such as finance, sales, operations, and supply chain. The site highlights financial services, healthcare, retail and CPG, and manufacturing.