05 / 15·PDPO, cross-border data, and private AI
What is enterprise private AI, compared with ChatGPT?
Three levels
Enterprise SaaS: a separate tenant and permissions inside the vendor cloud. Private cloud: you choose the region and network boundary. On-prem: the model and the data stay on your servers. All three still need a person to review output. The model does not send a contract by itself.
How that differs from a public chat tool
A public tool is built for personal productivity. Private AI is a company asset: who uploads, who can read, how long logs are kept, and whether chats train a model should all be writable policy.
Documents worth piloting first
Internal SOPs, masked invoice fields, and standard quote templates. Unmasked client identity and bank details wait for a later step.
Questions
- Does private AI mean fully offline?
- Not always. A private cloud is still online; you set the boundary and the permissions. Fully offline means on-prem, and someone has to run it.
- Can it replace a lawyer or accountant?
- No. Private AI drafts and classifies. Sending, booking, and signing still need a person.
This article is general information, not legal advice. PCPD guidance changes — check the latest version before you rely on it.
Want a first look at whether one real workflow touches personal data? Book a 1-hour business diagnosis.
Book a 1-hour diagnosis