Agent Audit

Paste any AI agent's URL, get a graded report in under a minute.

The audit gives you a single score, a verdict, and a list of strengths and weaknesses backed by actual transcripts. Save it to a Converra account to re-audit later and track improvement over time.

When to use it

• Before you ship — a sanity check on your own agent before it touches users
• After a prompt change — re-audit to see whether the change actually helped
• Comparing vendors — same eval set against three "AI customer support" agents you're evaluating
• Sharing a problem — the public report URL is a single link you can drop into Slack or a tweet

What you get

• One number, 0–100 — the score, anchored against an eval set
• A verdict — deploy, fix first, don't ship
• Strengths and weaknesses — what the agent did well, what it failed
• Real transcripts — every finding is backed by an actual conversation snippet
• Custom metrics — domain-specific rubric items (medical intake, returns, scheduling) layered on top of the default

Pricing

Tier	What you get
Free	1 audit per browser, lifetime
Pay-as-you-go	$9 per audit, no commitment
Pro	$299/mo, 15 audits + overage at $9 each
Enterprise	Custom — talk to sales

Next steps

• Quick start — submit your first URL
• Scoring — what's measured and how
• Re-auditing — before/after diffs