Trust starts before the pilot.

The public website is isolated from the SANDR product workspace. It uses an allowlisted promoted route manifest, restrictive browser security headers, request-size and method limits, rate-limited lead intake, release health reporting, and automated checks for private-file exposure. Unpublished preview routes remain outside public discovery and are no-indexed.

Each connected account must define the organisations involved, permitted data, purpose, access roles, retention, deletion or export path, service providers, incident contact, and any international transfer process required. The marketing form is not a channel for business files, staff reports, proofs, passwords, or sensitive personal data.

SANDR will identify the AI and hosting providers used for a pilot and document the settings and agreements that apply to customer data. We will not treat a broad website statement as a substitute for the pilot-specific data-processing agreement. Questions about AI training, retention, and provider access should be resolved in writing before onboarding.

Website enquiries are delivered only to the people and providers needed to respond. Pilot access must be role-scoped and reviewed with the customer. Credentials and secrets must not be submitted through the public form.

Security-related errors are recorded with the product version so the affected release can be contained or reversed. A pilot agreement will define the incident contact and notification process that fits the agreed data and service.

Email gm@sandr-ai.com with the subject “Security report”. Include the affected URL, a concise description, and safe reproduction steps. Do not access other people’s data, disrupt service, or publish sensitive detail before we have had a reasonable opportunity to investigate.