Your Agreements Are a Database You Can't Query — Hiral Shah, Docusign & Sean Sodha, NVIDIA
Docusign and NVIDIA use a small extraction model to preserve complex agreement tables as structured data, showing where specialized parsers can beat generic VLMs in agent pipelines.
Docusign says it processes **about 1 million agreements daily** across 1.9 million paying customers. Pricing, SLA, and rate-card tables often defeat line-oriented extraction, especially with merged or nested cells.
Use specialized extraction before retrieval when an agent must reason over table structure. The NeMo Retriever parse model preserves layout, reading order, and table relationships; Docusign reports **20× faster table extraction** than the open models it tested.
Docusign says it processes **about 1 million agreements daily** across 1.9 million paying customers. Pricing, SLA, and rate-card tables often defeat line-oriented extraction, especially with merged or nested cells. Use specialized extraction before retrieval when an agent must reason over table structure. The NeMo Retriever parse model preserves layout, reading order, and table relationships; Docusign reports **20× faster table extraction** than the open models it tested. The talk does not provide the underlying scores, model list, or production error rates. Docusign still uses a hybrid pipeline with OCR for other fields, while quantization and multi-token generation were described as future optimization paths.
This narrows model selection from choosing a general LLM to assigning a structure-aware model to the extraction stage before retrieval. The reported speedup suggests specialized parsing can matter at agreement scale, but the missing comparison details and continued OCR use make it evidence for a hybrid, workload-tested pipeline rather than a general open-model advantage. Quantization remains an unvalidated optimization here.