How agent-ready is Typeform?

Typeform's overall agentability score of 40/100 indicates moderate difficulty for autonomous AI agents attempting to parse and navigate its public web presence. The product performs adequately on machine readability (77/100) and control (72/100), meaning agents can generally parse the HTML structure and interact with forms and navigation elements. However, significant weaknesses appear in transparency (0/100), shadow UI avoidance (15/100), and defaults (18/100).

The transparency score of zero reflects a complete absence of machine-readable explanations for system behavior, confidence indicators, or audit trails that agents rely on to verify outputs and understand results. The low shadow UI avoidance score suggests heavy reliance on client-side rendering or interactive elements that may not expose their state to automated tools. Combined with minimal use of sensible defaults, these gaps mean agents must work significantly harder to extract information, distinguish authenticated from public content, and interpret results with confidence.

Performance varies across page types: documentation (52/100) is most agent-friendly, while pricing (33/100) presents the greatest obstacles. Chunking (41/100) and status communication (50/100) fall in the middle range, indicating room for improvement in how information is organized and how progress is communicated during multi-step processes.

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

• Add an activity log or 'why this result' feature that exposes system actions in both human-readable and JSON formats, enabling agents to track decisions and verify outputs.
• Surface confidence scores or 'verified vs. best-guess' labels on key outputs, particularly in API responses, so agents can assess reliability of information programmatically.
• Pre-fill form inputs with sensible defaults wherever reasonable values exist, and clearly mark which features and content are accessible without authentication versus requiring sign-in.
• Restructure headings to directly answer user questions (e.g., 'How do I cancel?' instead of 'Cancellation'), making content easier for agents to map to intent.
• Attach source references, input citations, or identifiers to generated outputs so agents can trace the provenance of information.
• Implement enumerated multi-step progress indicators or activity feeds for long-running tasks to give agents clear status visibility.