Validating

What is Agent-Readiness?

Agent-Readiness is an A–F grade summarising how well a product performs when AI agents are given real decision tasks about it. A means agents consistently include, rank highly, and recommend the product; F means agents systematically exclude or object to it.

The metrics behind the grade

  • inclusion_rate — % of agent runs where the product is shortlisted
  • mean_rank — average rank vs N competitors (lower is better)
  • hth_win_rate — % of head-to-head runs where the product wins
  • recommendation_rate — % of runs where the agent recommends the product
  • objection_density — average blocking objections per run (lower is better)

These are machine-observable metrics: they are measured from real agent behaviour, not from opinions. You define acceptance thresholds upfront, before any data is collected.

Why founders should care

The Advisory layer — an LLM recommending or omitting your product when asked — is already shaping buying decisions today. An Agent-Readiness grade tells you, with evidence, whether you show up when your customers ask AI for a recommendation, and what the machines object to if you don’t.