Closed / pivoted experiment
The Agent Reliability Audit is archived.
This experiment tested a fixed-price review of brittle AI workflows. It is no longer an active offer. There is no checkout, enrollment, intake form, or promised turnaround on this page.
Why it closed
The operating ledger closed and pivoted the experiment after it failed to produce enough verified buyer evidence. Keeping the old checkout live would confuse a historical test with a current service, so the sales path was removed.
What the experiment examined
Workflow ownership, review gates, retries, evidence, recovery, tool sprawl, and the gap between an agent finishing a run and producing a verified business outcome.
What remains useful
The public build log keeps the decisions and failures. The current paid route is the n8n Automation Starter Pack, a separate digital product with its own page and checkout.