Upsolve AI

Upsolve AI is an agent studio for data teams, founded in 2023, that lets them build analytics agents grounded in their own business context. Builders connect SQL data sources, encode definitions and validated SQL patterns, and deploy agents that end users can chat with to get answers and dashboards.

What It Is

Upsolve AI is an AI analytics platform with two sides: a builder studio for data teams and an end-user experience for conversational analytics. The homepage describes a three-layer context architecture covering structure (warehouse tables, SQL patterns, semantic models), meaning (definitions pulled from sources such as Notion, Slack and email) and trust (verified answers, golden sources, usage signals, lineage). It also offers embedded BI and GenBI so software companies can give their customers dashboards inside their own product.

How the Workflow Runs

Teams connect data through 30+ SQL database connectors, including Snowflake, BigQuery, Redshift, Postgres, Databricks and MySQL, and can import an existing dbt project; a pre-built semantic layer is not required. They then encode context, test and deploy, and evaluate and tune. Every conversation is traced end to end, from the user question through tool calls and generated SQL to the agent output. A built-in evaluation agent grades performance and context monitoring surfaces gaps.

Deployment Surfaces

Agents can be deployed to Slack, Teams, Claude, ChatGPT, Cursor, or embedded in a product via MCP or SDK, with the same context and guardrails on each surface. End users can also create interactive dashboards from a prompt using BI Canvases. For embedded use, the site says dashboards drop in via React or iFrame and respect row-level security.

Company Background

The About page says the founders previously worked at Palantir, and lists General Catalyst, Samsung Next, Soma Capital, Y Combinator and Twenty Two VC as backers.