Built Uses AWS to Power Document Intelligence
AWS ML Blog · rss · 2026-07-16
AWS highlights how Built Technologies built an AI document intelligence system on Amazon Bedrock and the AWS Intelligent Document Processing (IDP) Accelerator to support agentic products in real estate finance.
Real estate finance relies heavily on complex documents—such as draw packages, loan agreements, invoices, insurance certificates, and inspection reports—which vary greatly in format, length, and domain terminology. Built previously relied on 26 OCR and traditional ML-based processors. However, as use cases expanded to 250+ document types, millions of documents, and more complex reasoning needs, the legacy setup proved too rigid.
The article details the strategic shift from mere "extraction" to "understanding":
- Traditional OCR works for documents with explicit fields and stable layouts.
- Complex tasks, like identifying covenants in loan agreements, require contextual understanding, definitions, exceptions, and cross-paragraph reasoning.
- Agentic document workflows can classify document types, locate relevant sections, extract structured fields, retain evidence citations, and route low-confidence results to human experts.
Built designed these capabilities as a horizontally scalable foundation. They launched it first for commercial construction loan draw packages before expanding across the entire real estate finance lifecycle. The system compresses processes that previously took days into minutes while maintaining over a 95% confidence threshold for production readiness.
More from Infra
- Agent search bottlenecks are now about variance, not raw latency — rohanpaul_ai · 2026-07-22
- Gavin Baker argues Nvidia may be one of open source AI’s biggest supporters — GavinSBaker · 2026-07-22
- AI Power Demand Exposes US Energy Gap, Urging Shift from Scarcity to Abundance — bradneuberg · 2026-07-22
- Gavin Baker says Nvidia’s $630B figure would be system revenue, not all Nvidia’s — GavinSBaker · 2026-07-22
- A Firecracker-based platform says it can host 6,000 AI agents on one 256 GB server — maritime_sh · 2026-07-22
- Report says Nvidia could build 1,000 Vera Rubin racks a day, implying $630B quarterly at system level — GavinSBaker · 2026-07-22