Privacy

What we collect from this website

Only what you type into the contact or demo form: your name, email address, company, and whatever you tell us about your process. We also store the page you submitted from and your browser’s user-agent string, which help us tell a real enquiry from an automated one. There is no advertising tracker, no analytics cookie and no third-party script on these pages.

What we do with it

We reply to you. If you ticked the box asking to hear when ChatGPT support ships, we send you that one message. We do not sell it, we do not share it with advertisers, and we do not add you to a newsletter you did not ask for.

Workspace data

If you use Branchsolve, your workspace content — people, workflows, requests, decisions and attachments — is stored under your own organization and is processed on your behalf. A Branchsolve operator can create and suspend organizations and cannot read their contents. Operators can see limited service-health and included-AI allowance totals, never your prompts or generated responses. Reusable provider capabilities are stored as global declarative definitions and sanitized certification provenance; your connection, field mapping, workflow and acceptance remain inside your workspace and are never copied to another customer. Private automation proposals and their immutable revision history also remain inside your workspace. Export retains the bounded proposal text you supplied but removes creator identity and server-only binding proofs from those proposal records. Decisions are append-only: we cannot alter your audit trail, and neither can you.

Sign-in is handled by a managed authentication service. Branchsolve never stores your password.

When an AI-assisted authoring feature is enabled and you use it, Branchsolve sends only the content needed for that feature to a reviewed model provider. We record provider, model, purpose and billable token totals to manage the included allowance, but that accounting record stores no complete prompt, response or credential. AI authoring can pause at the allowance limit while saved workflows, requests, approvals and reports keep working.

How long we keep it

Website enquiries are kept while they are useful for the conversation you started, and archived rather than deleted so we have a record of what was said. Ask us to remove yours and we will. Workspace data is kept for as long as your organization is active.

Sub-processors

Google Cloud services host the application and its data. Branchsolve supports Anthropic and OpenAI as separately reviewed model-provider integrations, but each environment and workspace enables one provider and funding source explicitly. The selected provider processes authoring content only when the integration is enabled and you use an AI-assisted feature. A customer-provided credential is accepted once into server-side secret storage; the server retains only an opaque vault reference, while customer screens receive a bounded fingerprint and health metadata. We record sanitized usage and model provenance, never complete prompts, provider responses, credentials, or provider object IDs. Capability discovery returns only friendly published operation names, permitted fields, read/write risk, expected latency and whether your own healthy connection is compatible. It excludes raw provider schemas, tool identifiers, account references and provider responses. Automation authoring further limits the model to eligible logical operation IDs, friendly business fields, risk, latency and bounded mapping vocabulary. Organization identity, authority, approvers, connection/account selection, activation and execution are derived and checked by Branchsolve, not accepted from model output. Google’s mail service delivers the notification when you use a form on this site.

Your choices

Ask us for a copy of what we hold about you, ask us to correct it, or ask us to delete it. Get in touch and we will action it.