First audit intake
A focused export-first setup path for the $499 starter audit. This local UI shows the workflow only; it does not transmit files.
First audit
starter review
Setup required
for first audit
Intake scope
exports, concerns, docs, optional usage
Demo behavior
workflow only
Upload-first
Send conversation exports first
Drop conversation exports here
CSV, JSON, or helpdesk exports with AI author labels, timestamps, and conversation text. This local UI does not transmit files or connect to live helpdesk systems.
What to include in the audit package
The first audit is designed to work from scoped files and links before deeper integrations are justified.
Support exports
CSV, JSON, or helpdesk exports with AI author labels, timestamps, and conversation text.
Known concerns
Refund rules, escalation misses, churn-risk accounts, or answers your team already doubts.
Docs and policies
Public docs, help-center links, billing policy notes, and source material the AI should follow.
Scoped cost data
Optional weekly usage or spend summaries when you want cost anomalies included.
Represented sources for the sample digest
These are static demo rows showing the kind of sources a real audit package can reference.
| Source | Details | Status | Last run |
|---|---|---|---|
| Intercom export | CSV conversation export uploaded weekly | Connected | May 12, 7:30 AM |
| Zendesk AI replies | Ticket export with AI author labels | Needs mapping | May 11, 6:10 PM |
| Docs knowledge base | Public docs URLs and billing policy links | Connected | May 12, 7:42 AM |
| Usage ledger | Weekly AI support token and cost summary | Connected | May 12, 7:45 AM |
What comes back
Starter audit deliverables
Ranked incident digest
Plain-language summary of what went wrong, why it matters, and which items need action first.
Risky replies
Customer-facing AI answers that need correction, escalation, or human confirmation.
Doc gaps
Repeated confusion mapped into documentation fixes and source-material updates.
Owner action queue
Recommended fixes, impacted customers, owners, due dates, and follow-up context.
Human confirmation
High-risk findings should be checked before customer-facing follow-up.
Minimal intake
Send only the data needed for the digest, and redact sensitive details where practical.
Bounded claim
ReplyAudit surfaces support quality risk; it does not provide compliance certification.