Business Continuity Overview

Business Continuity Overview

1. In short

This document describes how Graine AI keeps its platform running and recovers from a failure, and what it does not yet do. It describes the approach, not the full continuity plan. Every statement is true as of 3 Nov 2026.

• Production runs in one AWS region, Mumbai. There is no second site or standby today.
• The primary database is a managed three-node replicated cluster with continuous backup by its provider.
• Graine has set recovery targets: a recovery time objective (RTO) of 24 hours and a recovery point objective (RPO) of 24 hours, with a maximum tolerable downtime of 72 hours. They are targets, not yet proven by a restore test (the first is Planned by 31 Dec 2026), and not a commitment. Graine does not quote an availability percentage.
• A tool that verifies restores is built and Being rolled out; no restore test has been run yet, and the first is Planned by 31 Dec 2026. The continuity and disaster-recovery plan was approved on 3 Nov 2026.

2. What is in place and what is not

Each Planned item shows its date. Until restore tests run, a backup has not been shown to restore by a repeatable test.

3. What this means in practice

Because there is one region and no standby, a failure of that region or of the servers the call platform runs on would stop calls, chat and the API until Graine restores service. Graine's recovery targets (service back within 24 hours, at most 24 hours of data lost) are not yet proven by a test, so Graine does not commit to a restoration time in advance. A customer who needs a committed recovery time for a time-critical process should ask Graine before relying on the platform for it.

The primary database does not depend on a single machine, and the provider backs it up continuously. A backup protects against loss of data. It does not keep calls running while the platform is down.

4. Vendors Graine depends on

Graine's recovery from a vendor outage depends on the vendor's own recovery. Most vendors include data processing terms in the terms Graine accepted; Graine does not yet have a signed or downloaded copy on file for any of them, and a few vendors publish none. Filing those copies, confirmed regions and review of the vendors' assurance reports are Planned.

5. How Graine tells you

When a disruption affects a customer, Graine's approach is to tell the customer by email, at the 24-hour contact on record where there is one, and to give an expected restoration time once it knows one. There is no public status page today; one is Planned by 31 Mar 2027. A personal-data breach follows the Incident and Breach Notification Commitments, which set a 24-hour notice. Graine has no staffed 24×7 operations centre; to report a problem at any hour, write to founders@graine.ai.

6. What you can do

1. Plan a fallback for a call that cannot be placed. Decide in advance whether the contact is retried later, moved to another channel or passed to a person. You can pause or cancel a campaign from the dashboard.
2. Make your webhook endpoint safe to receive a repeated or late event. Graine keeps a log of webhook deliveries for 30 days. After an outage on either side, reconcile from the call records in the dashboard or the API.
3. Keep your own copy of recordings and records you cannot afford to lose. Do not treat Graine as your only copy.
4. Keep a 24-hour contact current, so a notice reaches someone at any hour.
5. Keep your own telephony, carrier and provider settings documented, so you can reconfigure if you need to.

7. Contact

Related documents

• Security Overview
• Incident and Breach Notification Commitments
• Sub-processor List
• Data Retention and Deletion

Revision history