# Amandil — Full Content Reference > Amandil tells AI agents how healthcare works, process by process and rule by rule, with the source behind every answer. This file is the full text of amandil.ai in clean Markdown, for AI assistants and crawlers. Short version: https://amandil.ai/llms.txt --- ## Home URL: https://amandil.ai/ **Give your AI agents the healthcare knowledge they're missing.** Amandil tells your AI how healthcare works, process by process and rule by rule, with the source behind every answer. Live today: ~2,000 payer and provider processes · ~2,500 regulations linked to them · 9,000+ hospitals with CMS performance data · 22 operational domains. ### See it work Ask about a health system and get all of it, through one connection. Ask in plain language. Amandil returns the whole health system at once: every hospital, and every figure with its reporting period. - **The whole system.** Every hospital in a health system, not just the flagship. - **Every figure dated.** Each number carries its measurement period and how many hospitals reported it. - **One connection.** The work of finding, downloading and joining public files is already done. - **Its limits stated.** It says what the data can't show. ### Two sides of one coin Your platform knows your company. Amandil adds how healthcare works. Data platforms build context from your own tables and transactions. Amandil supplies the industry model: healthcare's processes, rules and how they connect. Example question: "Why are more of our imaging claims being denied?" - **Your company (what your data shows):** the imaging claim denial rate rose from 4% to 9%; many of the denied claims show authorization issues; patients are waiting longer between booking and their scans. Products of this kind include AWS Context, Databricks Genie Ontology, Google Cloud Knowledge Catalog, Microsoft Fabric IQ, Palantir Ontology and Snowflake Semantic Views. - **Healthcare (what Amandil adds):** an authorization may be valid only for specific service dates; a scheduling delay can push a scan beyond those dates; that can turn a scheduling problem into a payment problem. - **What the agent knows to check:** were the scans performed after their authorizations expired? Compare scan dates with authorization end dates. If that explains the denials, review how scheduling and authorization renewals work together. Amandil shows where to look. Your own data tests the explanation. An agent working in healthcare needs both. Amandil supplies healthcare context. Your connected systems supply the company-specific evidence. (Illustrative example with fictional data. Product names belong to their owners and do not imply affiliation.) ### Who builds on it - **Health systems.** The agents you build start from how each of your hospitals actually performs, not from a blank page. You get each hospital compared with national CMS measures, with the measurement period on every number; a ranked view of where AI and automation fit across your operation; and the processes and rules your agents need, in one connection. - **Payers.** Agents that work prior authorization and denials from the actual rule and process, with a source for every step. You get payer operations mapped process by process, from claims to utilization management; the regulations behind each process, linked to where they apply; and the same connection for the agents your teams build. - **Consultancies.** Weeks of discovery become a branded analysis. You get an account brief on a named health system or health plan before the first meeting; materials for executive working sessions, built from public benchmarks; and documents and decks produced in your own templates. - **Platforms and cloud.** Healthcare knowledge for the agents your customers build on your platform, and for the sellers who call on them. You get account playbooks for your sellers, from first meeting to renewal; healthcare knowledge for the AI agents your customers build on your platform; and AWS Marketplace and Databricks Marketplace listings, coming soon. ### Why not just search Your agent could rebuild this on every question. It shouldn't have to. - **One connection, nothing to build.** No downloading, cleaning and joining public files every time someone asks a question. - **The same data tomorrow.** Production agents need consistent, sourced figures, not a different method on every run. - **The whole system, not one hospital.** Every hospital in a health system rolled up, with the sites behind each number. Healthcare is first. Every regulated industry needs a model of how it runs. ### Questions - **What is Amandil?** Amandil is a working model of how the healthcare industry runs: the processes health systems and payers carry out, the regulations and standards behind each process, and public performance data for hospitals and health systems. AI agents and assistants connect to it so they answer from how healthcare actually works, with a source behind every answer. It was founded in 2026 by Sean Turner, formerly Chief Data Officer at Banner Health. - **What is the Operational Context Graph?** The Operational Context Graph, or OCG, is the name of Amandil's model. It links each healthcare process to the regulations that govern it, the measures that track it and the organizations that perform it, in a form AI tools can query. - **How does Amandil connect to our AI?** Through MCP, the standard way AI assistants and agents connect to outside tools, or through a REST API. Claude, Cursor and most agent frameworks support it. There's no data pipeline to build. - **Do you need our data?** No. Amandil is built from public sources: federal and state regulations, CMS data, industry standards and public announcements. We don't ingest patient data or connect to your systems. Agents you build can combine Amandil with your own data inside your environment. - **Can't our AI just search for this?** Yes. A capable assistant with web access and code tools can pull public files and work it out, and its method changes from run to run. Amandil does that work ahead of time and returns it through one connection, the same way each time, so agents running in production don't have to. - **How is this different from enterprise ontology products?** They're two sides of the same coin. AWS Context, Databricks Genie Ontology, Google Cloud's Knowledge Catalog, Microsoft Fabric IQ, Palantir's Ontology and Snowflake's Semantic Views each map your company from your own data. Amandil maps the industry your company operates in. An agent working in healthcare needs both. - **Where can we buy it?** Directly from us today. AWS Marketplace and Databricks Marketplace listings are coming soon. --- ## Founder URL: https://amandil.ai/founder.html **Sean Turner, Founder & CEO.** Twenty-five years leading data and AI inside health systems. Now building Amandil. - Previously: Chief Data Officer, Banner Health; System SVP, Data & Analytics, CommonSpirit Health; Interoperability & Population Health, Dignity Health. - Advisory, current: PwC Executive Advisor Network; Rackspace Healthcare Advisory Council; WebMD Ignite Executive Advisory Board. - Advisory, previous: AWS Healthcare Strategic Advisory Council. Amandil is building the working model of how healthcare runs, so the AI that hospitals, health plans and their partners build starts from how healthcare actually works. --- ## Trust & Security URL: https://amandil.ai/trust.html - **What Amandil is built from.** A model of how healthcare operates, built from public sources. It contains no personal data, and Amandil does not ingest customers' private data to build or operate it. - **No patient data.** Amandil does not connect to your EHR, claims systems or data warehouse. The Platform Terms prohibit submitting protected health information (PHI). If a service ever would involve PHI, a Business Associate Agreement is signed first. - **What Amandil holds.** Account and contact details for customer accounts, routine security and audit logs, and contact-form submissions from the website. - **Access.** OAuth 2.0 or bearer tokens, revocable immediately. Each organization's access is scoped to its own grant. - **Encryption.** In transit over HTTPS/TLS with HSTS; at rest on service providers' infrastructure. - **Retention and deletion.** Account records are deleted within 30 days of account closure or a verified request, except where law requires longer. Security and audit logs are retained. - **Reporting a security issue.** security@amandil.ai. Reports are acknowledged within 5 business days. --- ## Terminology - **Operational Context Graph (OCG)** is the name of Amandil's model of how healthcare runs. (Never "Operational Control Graph.") ## Contact - General: **hello@amandil.ai** - Partners: **partners@amandil.ai** - Security: **security@amandil.ai** - Web: https://amandil.ai/ ## Company Amandil Health, LLC. Founded 2026 by Sean Turner. Headquartered in the United States.