About AutomapAI: Semantic Normalization at Scale

AutomapAI™ is a semantic normalization platform and NLP-driven terminology server. It automates the mapping of unstructured EHR data, HL7 FHIR resources, and legacy clinical codes to global standards including SNOMEDCT, LOINC, RXNORM, ICD10CM, and USCDI.

The same platform can serve life-sciences organizations and researchers by supporting CDISC standards, OMOP data models, and more.

Take the data you have and map it to the data you need.

Aerial view of multiple overlapping highways surrounded by green trees and foliage, with several cars driving on the roads.

Benefits

  • Less manual effort. Automated matches give every mapping project a starting point, so subject matter experts spend their time on the decisions that need their expertise instead of searching from scratch.

  • Lower integration costs. Local codes, legacy standards, and unstructured text become standards-aligned data, without costly downstream cleanup.

  • Consistent, scalable mappings. Whether you're cleaning up legacy data, prepping for interoperability, or building clinical decision support, AutomapAI accelerates the process.

  • Ready for analytics and AI. Normalized data provides the reliability advanced analytics and AI depend on.

  • Within your security framework. AutomapAI operates entirely within your organization's internal security framework, so mapping is automated while security and auditability stay in your control.

  • ✺

    Terminology Crosswalk

    Develop concept maps from local terminology codes to the standards, or from one standard to another. Prepare initial maps in a fully automated way and finalize with light human curation.

  • ✺

    Semantic Standardization

    Start with FHIR resources, CCDA documents, or text strings containing medical content and normalize to USCDI or other terminology standards.

  • ✺

    NLP + Terminology Linking

    Start with clinical notes, extract relevant entities that you are interested in and link content to the terminologies you need. Can be used for concept indexing of journal articles, extraction of codable criteria from clinical trials write-ups, or extraction of clinical information from hospital notes.

AutomapAI review tool showing suggested matches with confidence scores

Real-world health data is complex, often existing as a mix of unstructured clinical narratives, scanned documents, local codes, and potentially legacy coded data. Much of this information is inaccurately coded or not coded at all, making interoperability effectively impossible.

AutomapAI converts complex, disparate data into standardized codes, delivering the accuracy and precision required to drive smarter patient screening, reliable querying, focused cohorting, and quality reporting across large or small healthcare data repositories.

How it works

AutomapAI uses a pipeline that takes coded or raw text input and produces matches in the desired output coding system, with a confidence score for each. It combines AI/ML, lexical matching, and semantic rules to identify and rank potential matches. When no match is confident enough, the term is marked as “provisional” so it can be subjected to followup review by humans or other processes.

AutomapAI can be used in these ways:

  • Task API. Send individual terms or batches of terms from your own systems and data pipelines and get matches back synchronously or asynchronously.

  • FHIR API. Send individual FHIR resources or bundles and the system augments “code” fields with additional codes as computed by the pipeline (and marked in a way to differentiate from original data). 

  • Via User Interface. Use the AutomapAI UI in a browser to load individual terms or batches of terms, evaluate suggestions, and review/manage results.

How we work with you

Choose the level of support that fits your team. Options can be combined.

  • Tooling access. We build a deployment with mapping pipelines to suit your use cases and make it available as a docker image that runs locally.  Use APIs programmatically or use the AutomapAI User Interface to explore and evaluate your data sets.

  • Consulting services. WCI's team can review, build, and maintain your mappings. When a map needs formal review and approval workflows, we deliver it through our Terminology Mapping Services, built on MT2.

Demonstrating Value

See a demonstration of our AutomapAI UI as it performs a mapping exercise on a set of content to derive USCDI mappings for conditions, procedures, lab results, and medications across SNOMEDCT, ICD10PCS, LOINC, and RXNORM.

Contact info@terminologyhub.com to sign up for an account to test in our reference environment.