-
01 Approvals Your queue, on the records being approved.
A manager opens the Approvals queue. Every item — a PO, an expense, a journal entry, a contract, a vendor onboarding — is the actual record being approved, not a copy in a workflow tool. The manager reads the diff — before and after, side effects, and which policy a rule flagged. One click ratifies and writes back to the source row in one transaction. Delegation is a record on the manager's account: time-bounded, audit-logged. No more amount drift between approval and posting, no more chasing an originator to find out what changed.
-
Approve the record
A PO approval writes to the PO. A journal approval writes to the journal. The amount cannot drift between approval and posting.
Diff and side effects
Each item shows before, after, and side effects of approval — which rows write, which notifications fire, which thresholds trip.
SoD on the row
Segregation of duties is enforced at the record, not at a workflow layer the ERP can't see.
Delegation as a record
Out-of-office delegation is time-bounded, audit-logged, and a queryable row — not a Slack message and a hope.
AI proposes route + flags
AI Builder proposes routing and surfaces policy-violation flags. The manager ratifies; agent_id is on every AI write.
-
02 Team rules Change a team rule, no ticket.
A Head of Sales wants quota credit to split 70/30 between closer and SDR. On the old stack, that was a CRM-admin ticket and three sprints. On oozmi, the Head opens admin, edits the credit-split rule as a record, and ratifies. The new split is live on the next closed-won; reps see updated quota attainment in real time; finance reads the new split for commission accrual. AI Builder proposes routing rules from the team's pattern; the manager reviews the diff and ratifies. Every change writes an audit row.
-
Team rule = record
Quota splits, routing, escalation thresholds, leave policy — every team rule is a row the manager edits in admin.
Live on next event
A rule change applies on the next closed-won, the next leave request, the next case escalation. No deploy cycle.
No ticket, no quarter
The bottleneck for a team-policy change is the manager, not the engineering queue. The org chart moves.
One-click revert
If the new rule misfires, revert is one click and one transaction. Risk-free experimentation on team policy.
-
03 Performance Team numbers from the rows the team wrote.
A manager opens the team dashboard at 09:00 Monday. Quota attainment, ticket SLA, time-to-close, AR aging — every number reads from the same rows the team wrote that week, not from a Friday export that already disagrees with the GL. The dashboard rolls up by team member, by deal stage, by SLA tier — same row, same number across the manager view, the IC's quota view, and the CFO's commission accrual. No reconciliation, no Friday-night CSV dump.
-
Read the live row
Dashboards query the same tables the team wrote that week. No export window, no reconciliation drift.
IC + manager + CFO agree
The rep sees the same closed-won total the manager sees, on the same row finance accrues commission against.
Rolls up, drills down
Group by team, by stage, by SLA, by region — the underlying row is always the source. No cube, no separate BI tool.
Anomalies surface live
AI flags numbers outside the team baseline as they post. The manager reviews; nothing slips to Friday.
Performance reviews drafted from the rows the team wrote — closed deals, tickets resolved, cycle times, peer feedback — with the manager owning every word that ships.
The substrate is the team's actual work. AI Builder drafts; the manager edits and ratifies. No new performance-management tool, no parallel system of record for ratings.
Production vs Pilot
- Approvals on the record — PO, journal, expense, contract, vendor
- Time-bounded delegation as a queryable record
- Team rule editing by the manager — quota splits, routing, escalation
- Live team dashboards from the same rows the team wrote
- audit_events on every approval, rule change, and team-record write
- AI-proposed approval routing and policy-violation flags
- AI-assisted performance review summaries from the team rows
- Anomaly detection on team KPIs (close rate, SLA, AR aging)
- Forecast-vs-actual variance flagged at the rep level
- Goal-setting with cascade across team rows
Where it fits
A manager runs the team with Ticketing , Workflows , CRM , AI Builder , and ERP — the same row across every module that wrote it. No second copy, no reconciliation step.
Often read alongside: Operations , People , and Front Office .
Bring one team rule you've been waiting on engineering for — we'll change it live.
A quota split your finance team won't recalculate mid-quarter, an approval threshold your CFO won't budge on without a forecast, a routing rule that adds a step every onboarding. Twenty-five minutes. We open admin against a sandbox of your team's data, change the rule, run a sample event through, and show what changes when the rule is a record the manager owns.