LegalTech Startup Transforms Desktop Legal Search Tool into Enterprise-Ready Cloud SaaS with Claude on Bedrock
At a glance
Legal discovery runs on speed and accuracy. For law firms and corporations managing litigation, the ability to search millions of documents and surface the right answer is a professional obligation, and missed or inaccurate results carry real legal consequences. Scaia.AI had proven NLP technology and more than 3,000 open-source users, but the product was a desktop application. Without a cloud-hosted, multi-tenant platform, enterprise clients, commercial subscriptions, and AWS Marketplace listing were all out of reach.
NMD designed and delivered a production-grade serverless platform on AWS, giving Scaia its first enterprise-ready commercial product: an API-accessible legal document intelligence service powered by Claude on Amazon Bedrock. The system ingests, processes, and semantically searches large legal document sets including PST email archives, returning natural language answers and document summaries at the API level.
Scaia.AI now operates a cloud-native platform ready for enterprise client onboarding and AWS Marketplace distribution, with two confirmed commercial use cases, eDiscovery and AI data-readiness assessment, and an active conversion path for its 3,000-plus open-source user base to a paid subscription tier.
Industry
Use Case
Intelligent Document Processing, Conversational AI / NLP, AI Platform Migration
Solution implemented
- Serverless document ingestion pipeline on AWS Lambda and Step Functions with S3 as the document store and SQS for event-driven processing
- PST email archive extraction and document chunking pipeline validated against the 5,600-plus email Enron dataset with DynamoDB tracking metadata across ZIP, PST, EML, and attachment levels
- Vector embeddings via Amazon Titan on Bedrock indexed in Amazon OpenSearch Service for hybrid semantic and keyword search
- Natural language Q&A, document summarization, and entity extraction via Claude on Amazon Bedrock at approximately $0.0002 per document
- Multi-tenant architecture with four subscription tiers, RBAC, project-scoped document isolation, and 33 API endpoints secured via API Gateway and Cognito
- Full Terraform IaC delivery with architecture diagrams, production runbook, and per-document cost model
The value equation
- Transformed a desktop-only LegalTech ISV into an enterprise-ready cloud SaaS product, opening law firm and corporate eDiscovery contracts
- Created a paid subscription path for 3,000-plus existing open-source users for the first time
- Delivered production AWS infrastructure required for AWS Marketplace listing and enterprise sales
- Enabled natural language Q&A, document summarization, and entity extraction across legal document sets via Claude on Bedrock
- Converted minimal pre-engagement AWS spend into a growing, production-grade cloud consumption footprint
Company Snapshot
Scaia.AI is a Houston-based LegalTech ISV that delivers AI-powered eDiscovery services and legal document analysis, enabling law firms and corporations to extract accurate answers from large volumes of unstructured case documentation.
Location
Houston, Texas, USA
Customer Situation
Scaia.AI had strong NLP technology and an established open-source user base, but the product was a desktop application. Enterprise law firms and corporations evaluating legal AI require cloud-hosted, multi-user platforms with API-level integration and multi-tenant document isolation. That is not a preference. It is a procurement requirement. Scaia’s desktop architecture made them ineligible for enterprise evaluation regardless of how capable the underlying technology was.
Moving from a monolithic Java application to a production cloud platform required expertise the team did not have: serverless compute, event-driven document pipelines, vector search, and multi-tenant access control. Every month without a cloud product was a month enterprise clients selected other options.
This is the defining failure point for LegalTech startups. Domain expertise and a working algorithm stall at the transition to enterprise-deployable product, and the gap is almost always cloud architecture and AWS program navigation, not the core technology.
NMD Solution
NMD reviewed the Scaia application and found a capable NLP engine constrained entirely by its desktop-bound, monolithic architecture with no cloud-native components. Building a commercial path required a full serverless redesign on AWS: Lambda and Step Functions for document processing, Bedrock for AI inference via Claude and Titan Embeddings, OpenSearch for retrieval, DynamoDB for metadata and state, and Terraform for repeatable deployment. NMD applied AWS POC funding to validate the architecture against the Enron dataset, then applied AWS Activate funding to build and deliver the full multi-tenant production system.
What We Delivered
Scaia.AI now operates a production-ready cloud-hosted legal document intelligence platform on AWS. Documents ingest via S3, process through a Lambda and Step Functions pipeline that extracts text, generates embeddings, and indexes content in OpenSearch for semantic and keyword search. Claude on Bedrock powers natural language Q&A, document summarization, and entity extraction across full document sets including PST email archives. The platform supports four subscription tiers, multi-tenant project isolation, and 33 API endpoints secured via Cognito and API Gateway. NMD delivered Terraform IaC, a production runbook, and a per-document cost model to support commercial pricing. Customer sign-off completed December 2025. Scaia is connecting its desktop application to the cloud API and pursuing AWS Marketplace listing as the primary distribution channel.
Ready to Transform Your AI Software Business?
See how New Math Data can transform your software business with AWS-powered innovation.