Context

Healthcare clinicians spend countless hours navigating complex software instead of focusing on patient care. A leading American healthcare technology company, bridges this gap through its Care Anywhere platform, connecting hospitals, clinicians, and home health providers. However, as the platform expanded, its paper-heavy documentation workflows became a bottleneck. Clinicians were bogged down translating multi-page forms, checkboxes, and vital signs from HomeCare HomeBase into standardized narratives and manually assigning complex CPT billing codes.

Challenges

Healthcare documentation remained one of the most time-consuming parts of the clinical workflow for the client. Patient visit notes generated through HomeCare HomeBase (HCHB) contained structured forms, checkboxes, clinician observations, vital signs, diagnoses, therapy notes, and billing information spread across multiple pages.

The existing process presented several operational challenges:

• Fragmented healthcare interoperability across multiple EMR systems created inconsistent patient documentation and delayed care coordination.
• Manual document interpretation required clinicians to review structured HCHB forms and convert them into comprehensive SOAP-format clinical documentation.
• Complex billing workflows required manual interpretation of patient records before assigning CPT billing codes, increasing turnaround time and the likelihood of coding inconsistencies.
• Limited scalability meant growing document volumes would require proportional increases in manual processing resources.

The client required an intelligent, cloud-native platform capable of automating clinical document extraction, report generation, and billing support while maintaining clinician oversight.

Searce Solution

Our team of solvers designed and implemented a production-oriented Generative AI healthcare document intelligence platform on AWS that automates the complete patient document processing lifecycle.

The solution combines Amazon Textract, Amazon Bedrock, AWS Lambda, Amazon API Gateway, Amazon S3, Amazon DynamoDB, and AWS ECS into a fully serverless, event-driven architecture capable of processing clinical documentation from ingestion through report generation.

The end-to-end workflow includes:

• Automated ingestion of patient visit PDFs through Amazon S3 or secure REST APIs
• Intelligent extraction of structured and unstructured clinical information using Amazon Textract
• AI-powered summarization using Anthropic Claude models on Amazon Bedrock
• Automated CPT billing code recommendation with clinical justification
• Generation of clinician-ready SOAP reports
• Secure storage of extracted data, reports, and metadata in Amazon S3 and Amazon DynamoDB
• Production-ready REST APIs for downstream healthcare integrations

The platform was designed to enable clinicians to review AI-generated outputs before downstream consumption, providing Human-in-the-Loop governance while significantly reducing repetitive documentation effort.

Business Impact

By pairing enterprise-grade Generative AI with a serverless data pipeline, the platform unlocked unprecedented scale while keeping clinicians firmly in control of patient care quality.

• 13x increase in clinical documentation and CPT-coding throughput — from ~150 to over 2,000 patient-visit documents per month — absorbed entirely within the original architecture, with no redesign required.
• Review headcount grew just 40% to support that 13x volume increase — lifting effective throughput per reviewer roughly 9.5x, and avoiding an estimated ~89% of the headcount growth a purely manual process would have required to keep pace. (Math: naive linear scaling would need ~13.3x headcount; actual growth was 1.4x → 1 − (1.4 ÷ 13.3) ≈ 89% avoided.)
• Directly reduces per-document labor cost of clinical documentation and billing-code generation, as automation — not incremental hiring — absorbs the majority of volume growth.
• Reduces downstream cost of coding errors and manual re-review across disparate EMR systems, by surfacing AI-generated CPT recommendations with transparent clinical justification rather than opaque, unexplained suggestions.
• Eliminates repetitive manual extraction work through 29 automated healthcare-specific queries, freeing clinical and coding staff to focus on judgment calls rather than data entry.
• Establishes a reusable, validated foundation — built on 52 annotated healthcare documents and exposed via six production APIs — positioning the platform for future healthcare interoperability and AI-assisted clinical workflows beyond this initial use case.

By transforming dense clinical forms into automated, intelligent workflows, Searce empowered this healthcare leader to dramatically slash administrative overhead. The resilient AWS foundation eliminates documentation bottlenecks today while clearing the runway for continuous, AI-driven clinical innovation tomorrow.