AtScale is a universal semantic layer that sits between cloud warehouses and the tools people use — Power BI, Tableau, Excel, Looker — and, through MCP, AI agents. Business logic is defined in SML, an open-source semantic modeling language, with versioning and CI/CD.
Product Overview
AtScale's angle is portability and determinism: one set of metric definitions, written in an open standard, executed in place on Snowflake, Databricks, BigQuery or Redshift without copying data, and exposed over SQL, MDX, DAX, Python and REST. MCP support lets conversational and agentic AI run analytics through the semantic model rather than generating SQL against raw tables. It's an enterprise data platform; RevOps benefits when revenue metrics are modeled there.
RevOps Jobs-to-Be-Done
- Same metric in Excel, Power BI and agents — Finance, RevOps and AI assistants query one definition of bookings or ARR from their own tools. KPI: Finance and the CEO stop presenting different numbers.
- Deterministic agent analytics — Agents query through MCP into the semantic model instead of writing their own SQL. KPI: Agent answers reproducible and auditable.
- Control warehouse cost — Aggregate management and query optimisation reduce compute spend for BI workloads. KPI: Lower cloud cost per dashboard query.
Key Features
- Open SML modeling: Code-first or visual semantic models in an open-source language, with versioning and CI/CD.
- Semantic query engine: Consistent metric resolution across SQL, MDX, DAX, Python and REST.
- MCP for AI agents: Lets agents execute analytics deterministically through semantic models.
- Query in place: Never stores or copies data; respects existing access controls.
- Metric store and governance: Governed definitions and metadata management.
How It Fits Your Stack
Primary system of record: Cloud data warehouse
Key integrations: Snowflake, Databricks, BigQuery, Redshift, Power BI, Tableau, Excel, Looker
Data flows: Translates queries from BI tools and agents through the semantic model into the warehouse; results return without data being copied.
Implementation & Ownership
- Time to first value: Weeks to months
- Implementation complexity: High
- Typical owners: Data platform team, BI team
Strongest where a large BI estate (especially Excel and Power BI) needs one set of definitions.
Pricing & Contracts
- Pricing model: Enterprise subscription
- Indicative range: Contact sales
- Free tier: No
Who It's Best For
Large enterprises with a cloud warehouse and many BI tools, where revenue metrics are owned by a central data team.
Good fit if:
- Finance runs on Excel and Power BI while others use Tableau or Looker
- You want semantic definitions in an open, portable format
- Agents must produce reproducible numbers
Probably not ideal if:
- You're a CRM-centric SMB or growth company
- You want self-serve pricing
Proof & Buyer Signals
Ratings: Leader, 2025 GigaOm Radar for Semantic Layers (as displayed by the vendor)
Pros
- Open SML avoids lock-in to one BI tool
- Broad protocol support including Excel and DAX
- Queries in place — no data copies
Cons
- Enterprise-only; no public pricing
- Heavy implementation for revenue use cases alone
- No GTM-specific identity resolution
Often Compared With
- Cube — Cube is developer-first with a free tier and its own BI; AtScale is enterprise and BI-tool-agnostic.
- dbt — dbt teams may prefer its built-in Semantic Layer; AtScale covers heavier multi-tool estates.
- Atlan — Atlan catalogs and governs context; AtScale computes the governed metrics.
Frequently Asked Questions
What is AtScale used for?
AtScale is a universal semantic layer that sits between cloud warehouses and the tools people use — Power BI, Tableau, Excel, Looker — and, through MCP, AI agents. Business logic is defined in SML, an open-source semantic modeling language, with versioning and CI/CD.
How much does AtScale cost?
AtScale pricing: Contact sales There is no free tier.
What does AtScale integrate with?
AtScale integrates with Snowflake, Databricks, BigQuery, Redshift, Power BI, Tableau, Excel, Looker.
Who is AtScale best for?
Large enterprises with a cloud warehouse and many BI tools, where revenue metrics are owned by a central data team.
What are the best AtScale alternatives?
The tools most often compared with AtScale are Cube, dbt, Atlan.
