Inference Analytics · agentic AI for healthcare
Build AI agents, put them to work across clinical, operational and revenue-cycle teams, and govern every action they take. One HIPAA-compliant core, deployed in your own cloud, on the models you choose.
Representative activity across the products.
Live insideUniversity of Chicago MedicineMayo Clinic
A different person in the building uses each one. All three share the same compliance boundary.
A no-code studio where health-system teams build RAG agents visually, ship them through a full lifecycle, and govern tokens and access across every model.
A multi-model AI workspace: each question goes to several leading models in parallel, and a supervisor model cross-checks them into one answer.
Denial and appeal automation and CPT coding checks, against payer policy libraries tracked daily, with a biller reviewing every draft.
Model-agnostic by design, and built on the same HIPAA-compliant core as every end-user application Inference Analytics ships.
Build once, run across OpenAI, Anthropic, Google, or internal models, connected only at the API level.
Govern access, usage, and cost per user, team, or agent, with a full audit trail for compliance review.
Design, test, stage and ship. Built for healthcare change control.
Close the loop between insight and agentic action, instead of handing off between data teams and the front line.
Every agent inherits the same compliance boundary. That is the reason health systems hand us backend tooling, not just apps.
80+ agents run on the same core today, which shows the platform doesn't need a rebuild for every new department.
Most programmes lose a year building governance before the first agent ships. That work is already done.
Typical elapsed time from engagement start to a first agent running in production, based on Inference Analytics deployments and customer-reported timelines for comparable approaches.
More than 80 agents run on Inference Analytics products. These illustrate the range a single compliant core has to cover, from the ward to the back office.
Specialty-specific agentic support, proving the core generalizes across very different clinical workflows.
Request a demo →Helps operational teams manage supply chain workflows on the same platform used for clinical agents.
Request a demo →Surfaces Epic tip sheets directly in clinical workflows, alongside the system of record rather than replacing it.
Request a demo →“UCM’s own IT organization builds agents directly in Dev Studio. It is infrastructure they build on, not a tool they consume.”
Inference Analytics makes three products for health systems. Dev Studio (InferAgents.ai) is where IT and informatics teams build, govern and ship AI agents across any model. Inferch is the multi-model AI workspace staff use every day. PayerAgents turns denied claims into cited appeals. All three share one HIPAA-compliant core, and more than 80 agents run in production today.
Yes. Dev Studio is the product and InferAgents.ai is where it lives. It is one of the three Inference Analytics products, alongside Inferch and PayerAgents.
They are model-agnostic. They orchestrate across OpenAI, Anthropic and Google models, and across a health system's own internal or self-hosted models, connected only at the API level. An agent is built once against the platform's interfaces, so the underlying model can be swapped without rebuilding the agent.
Protected health information stays inside the platform's compliance boundary. Model providers are reached only at the API level under a business associate agreement, data is encrypted in transit and at rest, access is governed per user, team and agent, and every agent action is written to an auditable log. Each agent inherits this core rather than implementing its own controls.
Yes. That is the intended use. University of Chicago Medicine's own IT organisation builds agents directly in Dev Studio. The platform is designed as infrastructure a health system builds on, not a closed vendor tool it only consumes.
Typically about three months from engagement start to a first agent running in production, compared with roughly eleven months on a generic cloud AI platform or sixteen months building in-house from scratch. The difference comes from the compliance core, governance and SDLC already being in place rather than being assembled per project.
See how our products fit into your model environment, compliance requirements, and existing agents.