How to Turn Agreement Data into Actionable Insights with Docusign IAM

Key takeaways

• Docusign Agreement Manager centralizes completed and historical agreements and uses AI to extract structured information such as parties, renewal dates, obligations, and custom fields.

• The Docusign CLI lets developers define, test, and deploy custom agreement types and AI-powered field extractions through a configuration-as-code workflow.

• The Agreement Manager API makes extracted agreement data available to downstream systems, including analytics platforms, CRM applications, embedded experiences, and automated workflows.

• The Docusign MCP Server enables compatible AI assistants to search and interpret Agreement Manager data using natural language.

Agreements hold some of the most valuable business information an organization has: pricing terms, renewal dates, obligations, and SLAs. 

But that information often remains buried in documents, forcing teams to review agreements manually, re-key data into other systems, and build workflows using information that may be incomplete or outdated.

This approach is slow, error-prone, and difficult to scale, especially as organizations adopt AI agents and automated workflows that depend on accurate, up-to-date data.

Docusign Intelligent Agreement Management turns completed agreements into structured data that developers can search, analyze, and use in downstream workflows and AI-powered applications. This article explains how Agreement Manager, the Docusign CLI, and the Agreement Manager API work together to turn agreement data into actionable insights.

A three-layered architecture for agreement intelligence

Docusign IAM addresses this challenge through a three-layered architecture that centralizes agreement storage, extracts structured data using AI, and makes that data accessible through APIs.

Note: IAM Toolkit is now known as the Docusign CLI.

Layer 1: Agreement Manager as a centralized repository

Agreement Manager serves as a centralized, AI-processed repository for every agreement that flows through Docusign. Every agreement sent and signed through your Docusign account, plus any historical agreements you upload, ends up in one place. 

You never need to pull from a third-party archive or shared drive. Because the agreements and their extracted fields are managed in one system, developers can trace the data back to its source document. 

Within Agreement Manager, you can see the information AI extracts across your agreements. Parties are automatically identified for each document. From the Show Insights dropdown, you can view agreements with renewals coming up in the next 12 months. Selecting an agreement opens its record, where you can review the AI-extracted fields for that document.

Layer 2: Docusign CLI for configuring Agreement Manager in bulk

As agreements flow into the Agreement Manager repository, AI surfaces structured metadata like pricing terms, obligations, and renewal dates. But IAM goes beyond generic document AI tools. If your business needs to extract a custom clause, a specific pricing term, or an obligation unique to your industry, you can define custom extractions in bulk using the Docusign CLI.

The CLI lets you configure and test custom extractions in a sandbox environment and promote them to production using a programmatic interface. The result is structured, trusted agreement data that reflects your business logic, not just out-of-the-box fields.

Layer 3: Agreement Manager API for integration

Once the fields have been extracted and tested, the Agreement Manager API makes them available to the tools your teams already use. Developers can then use the Agreement Manager API to bring that data into workflows, dashboards, and other business applications. 

Configuring custom AI extractions with the Docusign CLI 

The Docusign CLI provides a configuration-as-code experience for defining, testing, and deploying custom agreement types and field extractions. 

The process is straightforward: define the configuration in a JSON manifest, validate and refine it in a developer environment, and then deploy the validated configuration to production. 

Setting up your workspace

Start by installing the Docusign CLI and creating a local workspace for your Agreement Manager configuration:

# Install the Docusign CLI
npm install -g @docusign/agreement-cli

# Set up a workspace
ds scaffold -w my-workspace -p my-project -f agreement-manager

# Authenticate
ds auth login

The generated workspace includes three main JSON files in the configs folder:

• Standard catalog JSON: Definitions for the standard agreement types and fields provided with Docusign IAM

• Custom catalog JSON: Definitions for custom agreement types and fields already configured for your account

• Agreement Manager manifest JSON: The file where you define new custom agreement types and field extractions

You can retrieve the current catalog for your account and upload a completed configuration package from the CLI:

# Retrieve the current catalog
ds agm get catalog

# Upload a configuration package
ds agm upload

Defining custom agreement types and fields

In the Agreement Manager manifest, you define a custom agreement type by providing details like its display name, category, and AI definition. The AI definition describes when the model should classify a document as that agreement type.

"customAgreementTypes": [
    {
      "key": "C_MasterSupplyAgreement",
      "displayName": "Master Supply Agreement",
      "category": "BusinessServices",
      "customExtractionEnabled": true,
      "aiDefinition": "Master supply agreement governing the procurement of goods and services from a supplier, including insurance requirements, purchasing document references, delivery deadlines, and liability terms.",
      "fields": {
        "additionalStandardFields": [
          "Effective Date",
          "Initial Term Length",
          "Execution Date",
          "Termination for Convenience - Notice Period",
          "Termination for Cause - Notice Period",
          "Limitation of Liability Cap",
          "Payment Terms"
        ],
        "customFields": [
          "Commercial General Liability",
          "Cyber-liability and Data Breach",
          "Professional Liability",
          "Purchasing Document No.",
          "PD Date",
          "Related MSA Date",
          "Delivery Deadline",
          "Supplier Contact",
          "Has AI And Data Protection Terms"
        ]
      },
      "docs": [
        "AS- Powerbait MSA for Suppliers v20220515.docx",
        "Papermate Master Supply Agreement.docx",
        "Walmart_MSA_for_Suppliers_v20220515.docx",
        "Xfinity Master Supply Agreement.docx"
      ]
    }
  ],

You can then associate standard or custom fields with the agreement type. For example, suppose your organization needs to track whether its agreements include an AI and data protection addendum. You could define a custom field with:

• A category and field type

• Custom extraction enabled

• An AI definition describing the information to identify

• Examples showing how the addendum appears in agreements

Keeping these definitions in a local workspace makes the configuration easier to version, review, test, and reuse.

"fields": [
    {
      "key": "C_HasAIAndDataProtectionTerms",
      "displayName": "Has AI And Data Protection Terms",
      "category": "LegalAndComplianceCategory",
      "fieldType": "Boolean",
      "customExtractionEnabled": true,
      "aiDefinition": "If AI Services Addendum or similar AI and data protection terms are included in the agreement, set to true; otherwise, set to false",
      "examples": {
        "en": [
          {
            "exampleText": "This AI Services Addendum (“Addendum”) is incorporated into and made part of the Master Supply Agreement (“Agreement”) between the parties.",
            "confirmedValue": "true"
          }
        ]
      }
    },
]

Training the AI model

Custom agreement types require training documents that teach the AI model how to recognize the agreement type and extract its fields. Your training set should represent the different ways the relevant terms might appear.

For the AI and data protection field, include agreements with several versions of the addendum, along with examples where the addendum is not present.

Testing and validation

After you upload the manifest, the CLI creates the custom agreement types and fields and uploads the associated training and testing files into Agreement Manager.

Use testing documents that differ from your training set. You can generate a CSV test template containing the agreement and field names using the command ds agm test generate-test-template, then enter the expected value for each field in the resulting testing.csv file.

Agreement Name, Agreement ID, Agreement Type, field 1, field 2, ...
testdoc1.docx, 54c12508-xxxx-xxxx-xxxx-fc41c692d623, Master Supply Agreement, expected value 1, expected value 2, ...
testdoc2.docx, 2fb584d0-xxxx-xxxx-xxxx-74cd5826ba5e, Master Supply Agreement, expected value 1, expected value 2, ...
...

Running the tests compares those values with the AI-extracted results and returns an accuracy score for each field. If the AI and data protection field falls below your target, refine its definition, examples, or training documents and test again before deploying it to production.

Using agreement data in other applications

Once your custom extractions are defined and tested, the Agreement Manager API makes the resulting data available to other applications and workflows.

Searching agreement metadata

The API can query structured metadata extracted from agreements, including agreement types, parties, dates, provisions, and custom fields.

For example, you could search the custom AI and data protection field to identify which agreements contain the addendum and which do not. Because the query uses extracted metadata, teams do not need to open and review every document individually.

Importing historical agreements

The API can also support the ingestion of historical agreements into Agreement Manager. This allows organizations to bring existing agreement repositories into the same centralized system as completed agreements and apply AI extraction across a broader portfolio.

Building embedded agreement experiences

Developers can surface agreement information within the applications teams already use, such as CRM records, analytics dashboards, and AI-assisted interfaces. Users can access the relevant data without navigating to Agreement Manager separately.

In Tableau, for example, you could visualize the percentage of agreements that include an AI and data protection addendum. The dashboard could help legal or procurement teams identify gaps and decide which agreements may need updated terms.

Enabling AI-powered workflows

You can also use the Docusign MCP Server to make Agreement Manager data available through compatible AI assistants. After connecting it to Claude, for example, you could ask, “Are there any agreements in my account that don’t have an AI and data protection addendum?”

The assistant can search the custom field, identify the agreements missing the addendum, and present the results in natural language. Teams can then use those results to decide whether to renegotiate terms, send an updated agreement, or begin another follow-up process.

Bringing the components together

Docusign IAM gives developers a practical way to move agreement data beyond the document itself. Agreement Manager centralizes the agreements, the Docusign CLI lets you define and test custom extractions, and the Agreement Manager API makes the resulting data available to other systems.

Together, these tools can support use cases such as compliance reporting, renewal monitoring, embedded agreement experiences, and AI-assisted workflows.

Getting started

To start building with Docusign IAM, explore the Docusign CLI documentation for detailed setup instructions, API references, and implementation examples. The CLI provides everything you need to define custom extractions, test them in a sandbox environment, and deploy them to production.

From there, use the Agreement Manager API to bring the resulting agreement data into the workflows, dashboards, and applications your teams already use.