How agent-ready is Google Cloud?

Google Cloud's agentability audit returned an overall score of 36/100 on its homepage, with pricing pages scoring 30/100 and documentation at 28/100. These scores indicate that autonomous AI agents will encounter significant friction when attempting to parse information, complete workflows, or extract structured data from Google Cloud's public web presence.

The audit identified transparency (0/100), shadow UI avoidance (15/100), and defaults (23/100) as the weakest principles. Transparency scored zero due to the complete absence of confidence indicators, source citations, or drill-down summaries on key outputs. Shadow UI avoidance scored poorly because interactive elements lack proper semantic markup for agents to interpret. The defaults principle suffered from forms and configuration surfaces that leave inputs blank rather than pre-populating sensible starting values.

Relative strengths appeared in status communication (65/100) and machine readability (55/100), suggesting that Google Cloud provides reasonably clear feedback on system state and some baseline semantic structure. However, chunking (36/100), control (45/100), and clean handoffs (50/100) all scored below 50, indicating room for improvement in how information is organized, how users confirm critical actions, and how context transfers between interface states.

The following improvements would materially increase Google Cloud's agentability by addressing the weakest-scoring principles:

• Rewrite section headings as direct questions (e.g., "How do I cancel my subscription?" instead of "Cancellation") to improve content chunking and agent comprehension.
• Implement confidence indicators and certainty labels on API responses and UI outputs, with clear distinctions between verified data and inferred results.
• Add collapsible drill-downs to important results, displaying a one-line summary by default with expandable details available on demand.
• Include source references, input citations, or provenance links alongside generated or computed outputs so agents can trace data lineage.
• Pre-fill form inputs and configuration fields with sensible defaults wherever a reasonable starting value exists for the user's context.
• Add JSON-LD structured data markup using schema.org vocabularies to describe primary entities on each page, improving machine readability for agents and search systems.