On November 9, 2026, the Abu Dhabi government and OpenAI officially launched the Machine-Readable Government (MRG) initiative — a programme that translates selected regulations into a machine-readable format so AI can help officials handle applications for permits, licences and certificates. After the pilot at the Abu Dhabi Agriculture and Food Safety Authority (ADAFSA), the next wave spans five regulatory domains, including business and freelance licensing. Final decisions stay with people, and data stays in the UAE.
The headline on November 9: from pilot to five regulatory domains
Until today, the DGE × OpenAI collaboration ran in a single vertical — agriculture and food safety. The public launch of MRG marks the step from pilot into broader deployment. Per Gulf News and the Abu Dhabi Media Office, the near-term expansion covers five domains:
- building permits;
- business and freelance licensing;
- data governance;
- HR compliance;
- public procurement.
For entrepreneurs in the UAE, four of these five areas touch operational life directly — licensing, construction, HR compliance and procurement all sit at the heart of SME activity. The wider regulatory cycle is covered in our review of UAE business regulation in 2026.
What business gets: licensing, permits and HR
MRG does not change laws — it changes how they are applied. Every regulation that enters the machine-readable layer is decomposed into formal conditions that the AI matches against submitted documents. Ahmed Tamim Hisham Al Kuttab, chairman of the Department of Government Enablement of Abu Dhabi (DGE), put it plainly: "applying AI to old processes would only take us so far", calling MRG the "connective tissue" between the emirate's AI ambitions and real public services.
In practice, for a business that looks like this. The AI runs the routine parts of an application — formal conditions, cross-references, overlaps with other rules — in the background and prepares a brief with the specific clauses invoked. An official reads that brief and decides. Every assessment is traceable to the rule, which creates an audit trail. Partners emphasise that the system is not intended to take autonomous decisions on matters affecting human rights or access to essential services — those stay on a manual track.
The underlying company set-up process is covered in our guide on how to set up a company in the UAE step by step. MRG does not remove any of those steps — it is meant to compress the waiting time between them and make interpretations less dependent on who happens to review your file.
The Policy Adoption Engine: how it works
The technical core of MRG is the Policy Adoption Engine, developed by Inception42, a G42 subsidiary. According to Gulf News and specialist tech outlets, the engine turns individual policy clauses into sets of automated compliance checks: every rule becomes a machine condition with a clear link back to the originating clause. In parallel, MRG helps agencies spot overlaps and contradictions in regulations and test amendments before they are adopted.
OpenAI GPT inference runs on GPUs inside UAE data centres, under the previously announced UAE inference residency. Per the partners, the government retains ownership of all data going through the system, and that data will not be used to train OpenAI models. The design is described as fail-closed — if the pipeline is unavailable, the model simply does not return a decision.
Quotes and positioning
Sam Altman, OpenAI CEO, said at launch: "Making AI accessible for everyone requires transforming public sector delivery", adding that "Abu Dhabi has set a new standard for AI-enabled public services." For OpenAI the project is a rare whole-of-government reference and an argument in favour of sovereign infrastructure around GPT inference.
For Abu Dhabi, MRG is part of the Abu Dhabi Government Digital Strategy 2025-2027 worth AED 13 billion (about $3.5 billion), with a stated goal of becoming the world's first fully AI-native government by 2027. Per Gulf News, AI is already in use across 40 government entities in the emirate, all civil servants have been trained in AI, and the unified TAMM app covers more than 1,200 public services — the infrastructure base for scaling MRG is already there.
What it means for entrepreneurs
There are a few practical takeaways. First, timing: if the Policy Adoption Engine delivers in business licensing the same kind of time compression it delivered in food imports, parts of licensing procedures could move faster, while responsibility stays with the official. Second, predictability: anchoring every check to a specific clause of the rule narrows the grey zones in day-to-day interpretation. Third, cadence: one vertical pilot, then five, and only after that the full AI-native-government picture by 2027.
Specific go-live dates for MRG in the business and freelance licensing track have not yet been disclosed. We will follow the developments.
This material is for information only and does not constitute legal or investment advice. Specific procedures should be verified with licensed consultants and in current publications of DGE and the relevant UAE authorities.


