Built to fit how your MSP runs — with every AI rule on the record.
Custom roles enforced in the app and the database. AI switched on or off per company, with dispatch anywhere from manual to autonomous. Every rule that shapes what Corei can do is written to a registry a person can review — and learning suggestions stay pending until someone approves them.
Roles & permissions
Access that matches how your org is actually structured.
Custom roles from granular permissions
Build roles from individual permissions rather than picking from a fixed list.
Enforced in the app and the database
Permission checks aren't just UI — the database enforces the same boundaries.
Per-role and per-user layouts
Dashboards and workspaces can be tailored down to the individual, not just the role.
AI, your way
You decide how much autonomy AI gets — and where.
On or off, per company
Every AI feature can be switched on or off at the company level.
Manual to autonomous dispatch
Choose where each feature sits on the spectrum, from fully manual to fully autonomous.
Chat level & cleared work
Set Corei's chat level and exactly which work it's cleared to take on.
AI spend, tracked
See AI spend broken out by feature and by model.
Your operation, modeled
Corei is configured to your operation, not the other way around.
Boards per feature
Each feature gets its own board configuration matched to how that team works.
Complexity multipliers & skill overrides
Tune effort estimates and override skill scores where you know better than the model.
Branding, set once
Set your branding once and it's used everywhere — client portals, decks, and reports.
Clause rules, templates, retention
Contract clause rules, document templates, and data retention policies are all configurable.
Intelligence Control System
A registry, not a black box.
Every rule, recorded
Every rule that shapes AI behavior is written to a registry — nothing is a hidden default.
Suggestions stay pending
Learning suggestions from AI stay pending until a person reviews and approves them.
Reasoning traces to evidence
Corei's reasoning must trace to evidence — no unsupported claims stand unchallenged.
Security that starts with isolation
Isolation and access control, enforced below the application.
Tenant isolation, row by row
Every row is scoped to your company by database-level row security — not just application logic.
MFA across the app
Multi-factor enrollment and challenge are enforced across the whole application.
IP allowlisting, database-enforced
Enforced at the database level, with a 30-minute confirm-or-revert safeguard so a lockout can't be permanent.
Full detail on tenant isolation, MFA, and audit logging lives on the security architecture page.
The control map
Where AI helps, and who stays in control.
| Feature | Where AI improves it | Who stays in control |
|---|---|---|
| Autonomous tickets | Classifies, drafts and resolves routine tickets from historical patterns. | Confidence thresholds set by you; escalates anything uncertain. |
| Dispatch | Scores real skill from closed tickets, predicts difficulty, strips unsupported reasons. | Manual, assisted or autonomous; override anytime. |
| Phone intake | Transcribes, classifies and routes calls into the right queue in real time. | Reviews call outcomes; can require confirmation before ticket creation. |
| Corei | Answers questions and drafts work grounded in your own connected data. | Chat level and cleared work set per role; every answer traces to evidence. |
| Workspace | Surfaces the next best action and drafts routine responses. | Per-role and per-user layouts; suggestions stay optional. |
| Support chat | Handles common requests and hands off complex ones with context intact. | Dispatch mode set per company, from manual to autonomous. |
| Client data & QBRs | Assembles trends, recurring issues and talking points automatically. | Draft only; a person reviews and sends every deck. |
| Gauges | Tracks metrics continuously and flags confirmed anomalies. | Thresholds and alert routing configured by your team. |
| Quoting & procurement | Drafts line items, pricing and vendor options from past deals. | Nothing is sent to a customer or vendor without sign-off. |
| Billing & audits | Flags discrepancies between contracts, usage and invoices. | Corei never issues, voids, credits or collects. |
| Workflows & Radar | Confirms anomalies with evidence and triggers matching workflows. | People decide what happens next — Radar prepares, it does not act on its own. |
Try it — Illustrative demo
Set autonomy per feature, and review a learning suggestion.
This is a self-contained mock using sample data — no live accounts are connected.
Autonomy control panel · Illustrative demo data
Set the dispatch mode per feature. Every choice is recorded in the registry.
Learning suggestion pending approval
Corei suggests moving Support chat triage from Auto to Approval for billing-related intents, based on 14 recent escalations. This stays pending until a person approves or rejects it.
Intelligence Control
Configure the autonomy your operation is ready for.
Bring your roles, your risk tolerance, and your existing tools. We'll show you exactly what gets recorded in the registry and what still needs your sign-off.