{
  "id": 8945345,
  "title": "Docusign Processes 1 Million Contracts a Day on Small AI Models",
  "url": "https://urgent.news/2026/09/21/docusign-processes-1-million-contracts-a-day-on-small-ai-models",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-21T14:54:00.000Z",
  "source": {
    "name": "PYMNTS",
    "slug": "pymnts",
    "url": "https://www.pymnts.com/news/artificial-intelligence/2026/docusign-processes-1-million-contracts-a-day-on-small-ai-models/"
  },
  "original_language": "en",
  "account": "Docusign, a digital signature and contract management company, has implemented a cost-effective AI solution to process an astounding one million contracts daily. Initially, the company relied on large general-purpose AI models, but these proved costly and slow as the volume of contracts increased. In response, Docusign developed smaller, custom-trained models known as Iris for specific tasks, resulting in a 90% reduction in AI processing costs and an increase in throughput up to eightfold, according to Microsoft's case study published on September 15, 2023.\n\nDocusign's AI engine, Iris, analyzes each contract and extracts more than 50 critical facts, such as contract value, governing law, and renewal dates. Large customers often come with backlogs of tens of thousands of legacy contracts spread across 20 or more systems. When sending full contracts to powerful AI models, the per-call cost became a significant part of the infrastructure budget. Therefore, Docusign transformed the approach by treating different tasks based on their requirements.\n\nFor routine extraction, where the same data fields are extracted in the same manner across millions of documents, the company used smaller models trained in-house on Microsoft Foundry. For tasks requiring reasoning, such as understanding complex clauses or comparing terms across multiple documents, Docusign reserved its large frontiers models. By sending only relevant passages of a contract to the model instead of the entire document, Docusign achieved better results with a 7 percentage point accuracy increase compared to the previous system.\n\nThe cost per document fell dramatically, reaching 50 times lower than the previous system. This significant savings enabled Docusign to launch its next product line, which includes AI agents and an Agent Studio for automating contract work across various departments, including sales, HR, procurement, and legal. These agents rely on data extracted by the low-cost models, making the agents more affordable as well. SEI, a management consulting firm, utilized the system to extract obligations from master service agreements and connect them to Workday and HubSpot, streamlining the process of incorporating legal terms into new work orders.\n\nDespite many large U.S. enterprises adopting AI in data and technology functions, only a small percentage report fully paying back their investment within five to six years. However, at least half of the surveyed industries plan to increase their AI budgets next year, with none planning to cut back on AI spending.",
  "summary": "Even digital signatures and contracts have a need for artificial intelligence these days. Case in point: Docusign. It ran large general-purpose AI models on every contract it processed, and at more than 1 million agreements a day the cost and delays became a problem. The company swapped them for smaller models trained for specific jobs […] The post Docusign Processes 1 Million Contracts a Day on…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}