Autonomy is earned, dialed, and audited — never assumed.

Every agent runs inside the same framework: per-agent, per-property autonomy dials, human approval queues, deterministic compliance rails, a global kill switch, and an audit trail under every single action.

Approval-first by default — the AI proposes, you dispose
Dials per agent and per property, changeable any time
One kill switch stops every agent platform-wide, instantly

Book a live demo Create your company

Trust with receiptsEvery AI action logged with its grounding and outcome

Compliance in codeFair-housing and banned-language rails are deterministic

Reversible alwaysDial down or kill-switch without breaking the operation

What you get

Every claim below is a working screen in the product — open the demo and check.

The autonomy dial

Off → draft-for-approval → auto within bounds → autonomous, set per agent per property. Most operators start everything on approve and dial up what they’ve stopped editing.

Approval queues

Proposed messages and actions land in one queue with full context. Approve, edit, or reject — each choice teaching you where the dial belongs.

Deterministic guardrails

Fair-housing screening and banned collection language are enforced by code that runs on every message — not by hoping a prompt holds.

Bounds & policies

Pricing floors, plan limits, concession caps, escalation lists — agents operate inside the same org- and property-level policy system as your staff.

Global kill switch

One setting stops all agent activity platform-wide immediately. You’ll probably never use it; you’ll always know it’s there.

The audit trail

Human and AI actions in one reviewable log: who/what acted, on which record, grounded in what data, approved by whom. The license to operate, in table form.

Common questions from operators

What’s the recommended starting posture?

Everything on draft-for-approval. Watch the queue for two weeks; whatever you approve without edits is a candidate for dialing up — after-hours leasing first, because that’s where money leaks.

What can agents never do?

Cross your configured bounds (pricing floors, plan limits), use banned language (threats, legal posturing), bypass fair-housing screening, or act with the kill switch on. These are code-level constraints, not instructions.

Who can change the dials?

Role-based permissions govern dial changes, and every change is itself an audited action — so “who turned this up?” always has an answer.