Your agent's home screen shows three readiness gauges, each scored 0–100. They're diagnostic, not decorative — each one comes with a next-best-action telling you the single most useful thing to do next.
Answer readiness
How well your agent can answer support questions.
| Signal | What it measures |
|---|---|
| Knowledge base size | Whether you have enough content to draw on |
| AI search coverage | How much of your content has embeddings generated |
| FAQ coverage | Whether common questions are explicitly answered |
| Human-reviewed content | The share of your content that's curated rather than auto-generated |
| Open knowledge gaps | Questions visitors asked that your agent couldn't answer |
AI search coverage is the one to watch. Content without embeddings can still be found by keyword, but not by meaning — so a visitor phrasing a question differently from your article title won't find it. See Content isn't searchable.
Sales readiness
How well your agent can answer pre-purchase questions.
| Signal | What it measures |
|---|---|
| Pricing knowledge | Whether your plans and prices are in the knowledge base |
| Company story | Whether your agent knows who you are |
| Feature knowledge | Coverage of what your product actually does |
| Brand voice | Whether you've given the agent tone guidance |
| Sales routing | Whether sales-intent conversations are being routed |
If you don't sell through your site, a low sales score is fine to ignore.
Agent readiness
Whether the operational plumbing around your agent is set up — notification channels, and where escalations go.
How to use it
Don't chase 100. The gauges exist to point at the highest-leverage gap. In practice:
- Under 40 — your agent will frequently fail to answer. Add content.
- 40–70 — solid for common questions, patchy at the edges. Work the knowledge gaps list.
- Over 70 — healthy. Check back monthly as your product changes.
Scores are recalculated as your content changes. You can also fetch them programmatically from GET /api/sites/{siteId}/readiness.