Method
Method provides a platform for building structured, testable AI agent workflows, turning long prompts into versioned "Methods" with preconditions, checks, and recovery rules. The platform is used to convert dictated radiology notes into formatted reports using an example report's style, while keeping source quotes and validating drafts. Each run produces an independent record of results, allowing users to inspect changes, re-run workflows, and confirm repairs.
- Healthcare Technology
- Artificial Intelligence
- AI Agents
- Digital Health
- Developer Tools
- Software Only
Funding
Raised to date
$500KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Founder details are not available yet.
Product
Problem
Radiologists spend significant time manually drafting structured reports from dictated notes, a process that is slow, error-prone, and difficult to repeat across different cases. Long AI agent prompts that mix task instructions with accumulated rules become unstructured, hard to test, and cause agents to repeat failed steps, making automation unreliable and hard to improve.
Solution
Method provides a platform where users define a "Method"—a workflow file with instructions and automated checks for each step—that turns dictated notes into formatted radiology reports using an example report's headings and writing style. The system reads the example to extract a style guide, reads new notes, and drafts the report while keeping source quotes for each finding, then verifies the generated draft against the source notes before saving it for review. Each run produces an independent record of results, so users can inspect what was changed, re-run workflows, and confirm whether a repair worked. Method also supports broader agent workflows beyond radiology, such as research or data maintenance, and converts long prompts into structured, testable steps with preconditions, branches, and postconditions.
Target Audience
Primary customers are radiologists and radiology practices who need to draft structured reports quickly and reliably, as well as software developers and teams building AI agents for research, data maintenance, or other automated workflows who want inspectable, repeatable, and testable processes.
Features
- Converts dictated notes into a formatted report by extracting headings and style from a user-provided example report, never carrying over patient findings from the example
- Includes a multi-step pipeline that reads the example report, checks the extracted format, reads the new notes with source quotes, drafts the report, and validates the draft against the notes before saving
- Stores each workflow as a versioned file with retained values, preconditions, branches, approval gates, postconditions, and recovery rules, separating business rules from the prompt itself
- Provides run records with saved inputs, outputs, checks, errors, and timing, allowing users to inspect changed instructions and results
- Supports test-first workflows by keeping test inputs with the method and comparing returned values after changes to ensure rules behave correctly
- Includes a TypeScript SDK and dashboard for creating, running, and managing methods, compatible with agent runtimes like OpenAI models and coding agents