How agent-ready is DocuSign?

DocuSign's overall agentability score of 45/100 indicates moderate readiness for autonomous AI agent interaction. While the platform achieves perfect machine readability (100/100), meaning its HTML structure is well-formed and parseable, it falls short in helping agents understand what actions are available, what results mean, and how to navigate without authentication barriers.

The most significant gaps appear in transparency (0/100), defaults (18/100), and shadow UI avoidance (15/100). These low scores mean agents struggle to interpret the certainty of information they receive, cannot easily distinguish authenticated from public features, and encounter interactive elements that are difficult to detect or operate programmatically. Mid-range scores in chunking (42/100) and clean handoffs (50/100) suggest that content organization and navigation patterns are functional but leave room for improvement.

Stronger performance in control (73/100) and status (65/100) indicates that when agents do navigate the site, they receive reasonable feedback about available actions and system state. The documentation score of 40/100 is DocuSign's weakest surface, suggesting technical content would particularly benefit from better structure and transparency improvements.

To improve agentability, DocuSign should focus on the following concrete enhancements:

• Add confidence or verification indicators to key outputs, such as a 'verified vs. best-guess' label in the UI or confidence fields in API responses, so agents can assess information reliability.
• Expose a machine-readable activity log or 'why this result' feature that records system actions in both human-visible and JSON formats, enabling agents to trace decision paths.
• Clearly mark which features and content are accessible without authentication versus those requiring sign-in, allowing agents to plan navigation efficiently.
• Pre-fill form inputs with sensible defaults wherever reasonable values exist, reducing the cognitive load for agents attempting to complete workflows.
• Restructure documentation headings as direct questions (e.g., 'How do I cancel?') and lead each section with a 1–2 sentence answer before elaborating, making content easier to scan and extract.
• Attach source references, input citations, or identifiers to generated outputs so agents can verify information provenance and trace data lineage.