Every CRM, booking system and customer database stores personal data: names, phone numbers, email addresses and purchase history. In Georgia, its processing is governed by the Law on Personal Data Protection.
On 2 March 2026 the Personal Data Protection Service was abolished and its supervisory functions passed to the State Audit Office. Some businesses read that as a relaxation of the rules. It was not: the law and its obligations remain in force, and only the supervisor changed.
What changed in 2026
The Parliament of Georgia adopted the legislative package that abolished the Personal Data Protection Service and transferred its functions to the State Audit Office. Legal overviews of Georgian data protection law now name the State Audit Office as the supervisory authority.
For a business, this changes who supervises compliance. It does not change what the systems holding personal data must do.
Translate the obligations into system requirements
Legal texts describe principles; engineering teams need requirements. In system terms, the core expectations look like this:
- A recorded legal basis for each use of data, including when and how consent was given.
- Only the fields the business actually needs.
- The ability to find, correct, export or delete everything held about one person within the legal deadlines.
- Written agreements with processors such as hosting, email and analytics providers.
- Records of processing activities and a way to detect and report serious incidents quickly.
Sources and release notes: [2]
Build compliance into the system
When we build CRMs and internal systems, we treat these requirements as product features rather than paperwork. Access follows job roles, so a call-centre agent and a finance manager see different fields. An audit log shows who viewed, exported or changed each record.
Consent is stored as data, with its date, channel and the version of the privacy notice. Retention rules anonymise or delete records that are no longer needed, and a single screen can export or erase one person’s data on request.
AI features raise the stakes
When customer data reaches an AI model, for summarising calls or powering an assistant, new questions follow: which provider receives the data, where it is processed and whether it may be used for training. We recommend minimising personal data before it leaves your system and choosing deployments whose terms exclude training on your data.
A useful first step is an inventory of every system that holds personal data, what it holds, why, who can access it and who processes it for you. This article is general information, not legal advice.
Sources and release notes
Original analysis by AtmosGate, based on the sources below. Vendor claims and release status are attributed to their publishers.
