Thirty Minutes to Five
Co-Founder & CEO of CLRT
On 3 September the Dubai Land Department launched Initial Registration, one platform for project registration, transaction registration and escrow account management, with AI reading the Emirates ID, the passport and the sales contract so that nobody types them in. The release itself carries no numbers. The numbers came from Mustafa Al Rifai, the senior manager responsible for the system, speaking to WAM at the launch: data entry cut by up to 80 percent, a registration that took around 30 minutes now done in under five, more than 1,500 developers to be migrated in weekly batches. If you run a developer or a brokerage, you have probably spent the year being sold lead chatbots. Your regulator spent it automating the one step in your transaction that was never yours to speed up. The rest of the transaction is.
Read the release for what the platform actually does, because the mechanism matters more than the headline. A developer manages several companies through one account. The AI layer extracts the buyer's identity and the contract terms and populates the registration. Standard transactions that meet the requirements are eligible for approval on submission, which removes the counter read from the standard case. Above all of that sits Project 360, a consolidated view of each project's unit status, escrow accounts, financial data and early-warning indicators. Put together, the regulator has removed itself as the bottleneck. A Dubai off-plan transaction still runs from enquiry through reservation, internal approvals, contract drafting, buyer paperwork, escrow and handover, and the one step that used to be a fixed half hour of somebody else's time is now, by the official's stated figure, a few minutes of nobody's. Every remaining delay in the chain now belongs to the private firm.
The scale the platform absorbs explains why the regulator moved first. Data from the Dubai Land Department, as reported by Emirates 24/7 in July, put first-half 2026 transactions at AED 419.94 billion across 112,850 transactions, with sales of AED 286.44 billion through 86,000 deals. The department's own first-quarter release had already reported AED 252 billion, up 31 percent on the year, across 60,303 transactions. Inside the sales figure the shape of the registration load appears: completed property accounted for AED 146.69 billion across 27,160 transactions, while off-plan delivered slightly less value, AED 139.75 billion, across more than twice the number of deals, 58,840. Off-plan is where developer registration lives, and it is the half of the market that generates the paperwork. At that count, a half hour of counter time per registration is a department. The regulator did the arithmetic. Most developers have not done it on their own side of the file.
This is not the regulator's first model, either. In February 2025 the department reported more than 900,000 lease contracts registered in 2024, up 8 percent, and described its Smart Rent Index as using artificial intelligence to deliver standardised rental price assessments across every residential area. Under the rules it published, a landlord's rent increase applies only if the index confirms the property's eligibility. So the rent ceiling on close to a million contracts a year is already set by a model, not a negotiation. The pattern is consistent: the regulator points AI at high-volume, rule-bound, document-shaped work with a clear system of record: registering, pricing, checking. It has not pointed AI at persuasion. The private side of the sector has spent its budget on the opposite: the chatbot on the listing, the automated follow-up, the lead qualifier. Neither is wrong. But one party is now measurably faster than the other at the step they share.
Which brings us to the number itself. Up to 80 percent is an official's statement at a launch, carried by the state news agency, and it deserves to be read as one: a stated figure, not an audited measurement of your transaction. The temptation is to quote it, in a board deck or an investor update, as evidence that the sector's cycle time has fallen. It has not, and not for you. The regulator's step was perhaps 30 minutes of a process that takes weeks. What the launch claim should trigger is a measurement, not a citation: how long a transaction takes in your firm from reservation to registered title, and what share of that time sits inside the regulator against inside your own approvals, your contract desk and your buyer's missing document. Most firms cannot produce that number in a day, and that inability is the finding. A launch figure is a prompt to test your own clock. Until you have, it is colour.
The uncomfortable arithmetic is that a firm which is slower than its regulator is also, in most cases, less instrumented than its regulator. Project 360 gives the department a consolidated, early-warning view of every registered project. It is entirely possible that the regulator can see a developer's escrow position and unit status before that developer's own management can. That is the real signal in the launch, and it has nothing to do with chatbots. The question for a principal is where, between enquiry and handover, the next fixed half hour of somebody's time is hiding, whether it is document-shaped and rule-bound enough to be worth automating, and what a wrong output costs at that node when the regulator's counter is no longer there to catch it. Those are judgment questions, specific to one firm, and the regulator has just shown what answering them well looks like.
Your regulator spent the year automating the one step you could never speed up. Everything left in the chain is yours.
A deeper dive
The mechanism behind the official's figure is worth understanding because it shows where the speed came from and where it cannot come from. Initial Registration is fast for two reasons that compound: the identity and contract fields are extracted by a model rather than typed, and a transaction that meets the published requirements is approved by rule rather than by a reader. Both depend on a clean system of record, a fixed definition of compliant, and enough volume that the boring cases dominate. A private developer's back office has none of those properties by default. Its approvals route through people whose availability is the real clock. Its sales contract is drafted from a template that has been edited by hand for a decade. Its buyer's paperwork arrives in fragments over WhatsApp. None of that is document reading, and none of it becomes faster because the counter at the end got faster. It becomes visible. A fast node at the end of a chain exposes the slow nodes before it, and the first honest map of a developer's cycle time usually shows the regulator as a rounding error.
The second-order trap is quieter and more expensive. For years the half hour at the registration counter was not only a delay, it was a check. A person read the file, and a mis-keyed unit number, a price that did not match the reservation, a buyer whose name was spelt two ways, were caught there by somebody else, at no cost to the developer. Automatic processing of compliant standard transactions means a file that is compliant in form passes in minutes, whatever it contains. The error does not disappear; it moves downstream, to escrow, to handover, to a dispute, where it is dearer to unwind and where the early-warning indicators in Project 360 will show it to the regulator first. A firm that responds to the launch by pushing files through faster, without building its own verification where the counter used to be, has traded a visible delay for an invisible liability. The regulator automated its check and kept its oversight. The developer who assumes the regulator's speed is now theirs has automated nothing and lost a check. The right response to a fast regulator is a firm that knows its own cycle time to the day and owns verification at every node the regulator no longer reads.
Work with CLRT
CLRT does exactly this mapping for developers, brokerages and the firms that serve them: the honest cycle-time map of one transaction, the nodes where a model belongs and the nodes where a person's judgment is the product, and the verification layer that has to exist wherever a counter check used to. The regulator has shown what pointing AI at the right node looks like. The question is where that node sits inside your firm, and the CLRT Ascent diagnostic at ascent.clrtstudio.com is where we find it. Talk to CLRT before the next launch figure becomes a slide in your deck.
Vishal Sachar is the Co-Founder and CEO of CLRT, where he helps UAE businesses make sense of applied agentic AI and put it to work. He writes on agentic systems, AI governance, and the economics of automation. Reach him at vishal@clrtstudio.com or on LinkedIn.
