Grounded drafts
Draft only from the approved policies and knowledge your team supplies.
Private pilot for property and operations teams
Answer routine inbox questions. Escalate the ones that need judgment.
Grounded Inbox prepares review-ready answers from approved knowledge, while emergencies, rent hardship, policy gaps, and uncertainty always route to a person. Every send still needs human approval.
It brings AI email triage and human-in-the-loop approval into one controlled workflow—so routine questions move quickly and sensitive cases stay with people.
We are onboarding a limited number of teams. No inbox connection is required to explore the demo.
How it works
Automation where it is safe. Human judgment where it matters.
Deterministic escalation
Route emergencies, sensitive categories, uncertainty, and missing knowledge to people.
Human-controlled send
Nothing is sent until a person explicitly approves the reply.
Who it helps
Built for teams whose inbox is part of the customer experience.
Property management
Prepare review-ready replies to routine leasing and resident questions, including pet policies, parking, application requirements, and lease-renewal terms.
The team stays in control of every reply before it is sent.
Maintenance emergencies, rent hardship, fee disputes, and policy gaps route straight to a person instead of being improvised by AI.
Support & operations
Keep routine billing procedures, invoice-copy requests, and account-status questions moving from approved internal guidance.
Give reviewers a grounded draft instead of a blank compose window.
Disputed charges, account-closure requests, sensitive issues, and unclear cases are escalated for human judgment.
Customer-facing teams
Use one controlled workflow for common questions arriving by email, SMS, WhatsApp, or manual intake.
Teams can respond consistently when policy coverage is clear and current.
Urgent requests, sensitive topics, missing knowledge, and low-confidence messages stop for review rather than receiving an automated answer.
Property managers, support teams, and operations teams often handle the same policy questions while trying not to miss the few messages that require judgment. Grounded Inbox prepares replies from approved knowledge for routine questions such as lease terms, pet policies, service requests, billing procedures, and account information. When a message is urgent, sensitive, unclear, or outside approved guidance, it is escalated instead. Teams respond faster without improvising policy answers or losing the human approval step before send.
See the boundaries in action
Explore a safe, preloaded inbox.
Walk through suggested replies, escalations, unsafe-instruction handling, and the human approval step using sample data.
Questions, answered
Automation without giving up control.
How does AI email triage work?
Grounded Inbox normalizes an incoming message and checks it against the approved policy and process knowledge your team maintains. When the support is sufficient, it prepares a draft for review; when it is not, the message is escalated instead.
What happens when a message is escalated?
Grounded Inbox stops before preparing or sending a reply. A person can review the message, resolve it, or prepare a human reply through the same controlled process.
Does Grounded Inbox send emails automatically?
No. Every reply requires explicit human approval before it can be sent. The approval step remains visible and auditable, even when the draft is well grounded.
Which inboxes does Grounded Inbox connect to?
The private pilot currently connects Gmail and Google Workspace inboxes. It is designed to bring email, SMS, WhatsApp, and manually added messages into one review experience as those channels are introduced.
Private pilot
Bring your real inbox into a controlled workflow.
We will walk through your policy knowledge, escalation boundaries, and approval process before inviting your Gmail or Google Workspace account.
Request pilot access