# Agentic Data Intelligence - Full Context > Last updated: 2026-07-01 > The control plane for AI data analytics. Automatically generated machine-readable semantics and a contract that enforce the behavior of AI agents during analysis. This is the extended context file for LLMs. For a concise summary, see [/llms.txt](/llms.txt). --- ## Company Overview Agentic Data Intelligence SA is a Swiss-headquartered startup building the semantics and control layer for AI data analytics - trusted, governed, and ready for production. ADI was originally developed to deal with some of the challenges we ourselves face when dealing with some of the most complex data possible: on-chain blockchain data. The modern data stack has never been more capable, but most organisations still struggle to turn raw data into answers they can trust. Copilots write SQL quickly from a natural language question but cannot tell you whether the answer is right. Two dashboards, same question, different numbers. That is not a tooling problem - it is a meaning and trust problem. Business meaning is scattered across SQL, dashboards, documentation, and people's heads. No layer in the stack owns it. Every tool downstream is working with incomplete context. Agentic Data Intelligence solves this - whether you are an individual using agentic workflows, or an enterprise using cloud data warehouses. Our AI agents understand what your data means before they try to do anything with it. They build an explicit model of your business semantics, then work within that model to answer questions and enforce correctness by design. You stay in control: approving changes to the semantics and contract. The human role shifts from writing routine SQL to defining what matters to the business. Headquartered in Switzerland and backed by leading venture capital firms. Working with early partners to bring agentic data engineering and analytics into production. --- ## Products ADI is data and platform agnostic. The Starter Edition is built for coding agents and IDEs; the Enterprise Edition is built for cloud data warehouses. More integrations are being added in the coming months. ### Starter Edition **Built for:** Individual developers, data engineers, and analysts wanting to add semantic and control functionality to agents and agentic workflows that conduct data analytics. **Focus:** Designed to work with coding agents and IDEs. **Delivery:** SKILL.md files, plugins, and MCP Servers. Domain Packages for different data types to be released in the future. **Features:** Basic semantics and rules. Standard authentication and security. **Integrations available now:** Claude Code, OpenClaw, Cursor, Manus. **Coming soon:** Codex. ### Enterprise Edition **Built for:** Startups and enterprises that use copilots and agents for data analytics, and want to add a semantic and control layer to improve governance, accuracy and consistency. **Focus:** Designed to work with cloud data and analytics platforms. **Delivery:** Native apps that extend (not replace) existing catalog and governance, ensuring data stays in your warehouse. **Features:** Full and complex semantics, rules, and analytical processing. Strictly enforced authentication, security, and data governance. **Integrations available now:** Snowflake, Databricks, dbt. **Coming soon:** Google BigQuery, Amazon Redshift. --- ## Capabilities ### Understands Your Data ADI creates its own semantic understanding of your raw data and generates a machine-readable semantics summary and contract (rules) for agents and copilots to use. It models your business explicitly - entities, events, metrics, grain, and constraints - instead of guessing from schemas or SQL. Because meaning is machine-readable, the system knows which metrics can be computed, how they can be aggregated, and which questions are invalid. ### Managing Change and Drift in Semantics One of the biggest challenges with semantics is that they rapidly go out of date. With ADI you can easily update, and see past versions of, both the semantics and the contract. Changes are explicit and reviewed, not silent adaptation - updates can be approved, tested, and rolled back so pipelines and analytics evolve with the business while remaining stable, auditable, and trustworthy. ### Minimal Footprint and Information Transfer ADI never sees your underlying data. The contract is generated from schema and context, but the analysis is run locally by you. ### Domain Agnostic Works with any data, whether your own or from marketplaces: product, sales, financial, market, and even blockchain data. ### Works Where You Conduct Analytics ADI is a thin layer that works with coding agents, IDEs, or data warehouses. You ask questions as normal and ADI's control plane is working quietly in the background. ### Enforce Behavior of Agents By their nature, the LLMs behind agents are non-deterministic - the same question can result in two different answers. ADI creates a contract that forces agents to use the data and generate SQL in the way that is expected. ### Tech Stack Built with Anthropic APIs, Pydantic AI, Pinecone, and Supabase. Starter edition also uses FastMCP and the Agent Skills framework. ADI supports the Open Semantic Interchange (OSI) initiative and is working towards two-way compatibility with OSI standards in future versions. --- ## Current Status Accepting first partners. We are accepting users to the waitlist for our limited alpha release; request a place via the early access page at https://agentic-data.com/access. Looking for partners across industries who want to bring agentic data engineering and analytics into production. --- ## Content and Resources - **Company blog:** https://agentic-data.com/resources (product updates and company news) - **Research and technology (Substack):** https://agenticdataintelligence.substack.com (long-form research exploring AI, data engineering, and analytics) - **Newsletter:** available via the resources page --- ## Frequently Asked Questions **What is Agentic Data Intelligence?** Agentic Data Intelligence (ADI) is a thin layer that sits between agents and data, to automatically generate machine-readable semantics and a contract that can be used to enforce behavior of co-pilots and agents conducting data analytics, and thereby to increase governance, accuracy and consistency. See our Introducing Agentic Data Intelligence blog series for more information. **What is the difference between the Starter and Enterprise editions?** The Starter edition is aimed at individual developers, engineers and analysts who want to enhance data and analytics functionality to their agents and agentic workflows. The Enterprise edition is aimed at startups and companies who use cloud data warehouses and want a more agentic approach to business intelligence without sacrificing governance or business meaning. **What integrations are there?** The Starter edition is designed to work with Coding Agents and IDEs like Claude Code, OpenClaw, Cursor, and Manus, with Codex coming soon. The Enterprise edition currently works with Snowflake, Databricks, and dbt (and MetricFlow), with Google BigQuery and Amazon Redshift coming soon. **What powers ADI?** ADI is built using leading AI frameworks, libraries, tools and LLMs including Anthropic APIs, Pydantic AI, Pinecone, and Supabase. Our Starter edition also uses FastMCP and the Agent Skills framework. We support the Open Semantic Interchange initiative and are working towards two-way compatibility with OSI standards in future versions. **How do I get started?** We are accepting users to the waitlist for our limited alpha release. You can request a place on our early access page at https://agentic-data.com/access. --- ## AI Practices - AI used for: data analysis and pattern recognition, automated decision support systems, natural language processing, predictive analytics - Does not use customer data to train models without explicit consent - Human oversight maintained in critical decision-making - Full AI notice: https://agentic-data.com/ai-notice --- ## Legal and Contact **Company:** Agentic Data Intelligence SA **Address:** Esplanade de Pont-Rouge 9A, 1212 Grand-Lancy, Lancy, Switzerland **Founded:** 2021 **Governing law:** Swiss law **Jurisdiction:** Ordinary courts of Geneva, Switzerland **EU GDPR representative:** Prighter (https://prighter.com/q/15639787740) **Contact:** - Privacy: privacy@agentic-data.com - Legal: legal@agentic-data.com **Social:** - GitHub: https://github.com/agentic-data-intelligence - LinkedIn: https://www.linkedin.com/company/agenticdata - X/Twitter: https://x.com/agentic_data (@agentic_data) **Legal documents:** - Privacy Policy: https://agentic-data.com/privacy - Terms of Service: https://agentic-data.com/terms - Cookies Notice: https://agentic-data.com/cookies - AI Notice: https://agentic-data.com/ai-notice