Product
Corpus
The operator workspace for reviewable AI
Ingest sales and support conversations, evaluate intent and risk, and produce AI reply drafts your team can approve, edit, or hand off—with a clear trace of every decision.
01 — Workspace
One place to run customer conversations
Operators work from a single inbox that surfaces conversation state, contact context, and AI status without leaving the thread.
- Conversation list with pending review, handoff, and safety labels
- Thread view with customer history and live updates
- Contact and attribute context beside the conversation
- AI insight panel for draft readiness and ownership
Screenshot
Operator workspace
Inbox · thread · contact · AI panel
02 — Drafts
AI drafts stay under human control
Corpus generates reply drafts for sales and support. Nothing customer-facing goes out until policy allows—or an operator acts.
- Approve, edit, regenerate, or reject each draft
- Safety summary beside the proposed reply
- Risk and uncertainty cues before you commit
- Read-only drafts until review or explicit automation policy
Screenshot
AI draft review
Approve · edit · regenerate · reject
03 — Trace
See how every draft was decided
Operators get a clean timeline: ingestion, supervisor routing, safety findings, draft state, and review status—without digging through logs.
- Supervisor intent and risk decisions on the record
- Input and output safety findings surfaced clearly
- Draft generation and review status in one drawer
- Timestamps and status chips for audit-ready clarity
Screenshot
AI trace drawer
Ingest → evaluate → safety → draft → review
04 — Handoff
Pause, escalate, keep the context
When AI is uncertain or the case is sensitive, ownership moves cleanly between sales, support, admin, and human review.
- Clear pause and escalation states on the conversation
- Ownership that survives team-to-team handoffs
- Context preserved across routing boundaries
- No orphaned threads when automation steps back
Screenshot
Ownership and handoff
Pending review · human handoff · conversation owner
05 — Admin
Account controls for how AI may act
Admins configure teams, labels, attributes, channels, and review policy so autonomy is explicit—not accidental.
- Team and channel configuration for operations
- Autonomy and review-required policy toggles
- Escalation conditions and safety thresholds
- Account-level settings that bind every conversation
Screenshot
Account admin controls
Teams · labels · autonomy · review workflow
06 — Superadmin
Operations that keep accounts healthy
Internal operators provision accounts, watch integration sync, recover failed jobs, and inspect AI decision audit status.
- Account provisioning and integration sync status
- Failed jobs and dead-letter recovery tools
- AI operation audit visibility across accounts
- Dense, filterable views built for ops—not marketing
Screenshot
Superadmin observability
Provisioning · sync · failed jobs · audit
Integrations
Your channels connect conversations. Corpus is the operating layer.
Corpus ingests and syncs messages, contacts, labels, and conversation metadata from the tools you authorize. Integrations bring the signal in—Corpus owns review workflows, safety gates, traces, and account controls.
Enterprise Solutions