We at Viewit have been dabbling with AI in PropTech for a while, since 2023 when OpenAI first launched their APIs. We built an AI virtual agent trained on 8 years of Dubai Land Department transaction data that can answer any question to do with Dubai real estate, backed by DLD data. We integrated AI into our CRM where, with the click of a button, you get an SEO-optimised description of the listing you want to post on viewit.ae or viewit.pk. We are testing (and soon to launch) an AI video editor where you can drop in any video tour of a property and it will identify each room, annotate the video, slow down panning, speed up walking, and eventually narrate the video using AI. Lastly, we are working on an AI recommendation engine that learns as you scroll our video feed on our websites or apps and will allow you to successfully scroll into your next home. These are good products and they enhance current workflows. But is this the future of PropTech? Let’s put our prediction hats on and see how these trends play out.
PropTech investment hit $16.7 billion in 2025, up 68% from the year before, with AI native companies accounting for around $4.5 billion of that total and growing at nearly double the rate of non-AI players. The deals tell the story: EliseAI raised $250 million at a $2.25 billion valuation for AI-powered conversational agents that handle leasing inquiries, scheduling, and resident communications. Bedrock Robotics, founded by former Waymo engineers, raised $270 million in February 2026 for AI-driven construction automation. Juniper Square closed a $130 million Series D to automate commercial real estate syndication (a group of investors buying commercial real estate together). Three of the four newest PropTech unicorns are AI-native, all minted since mid-2024. The capital is chasing companies that can make entire real estate functions autonomous and own the whole value chain. Y combinator now only interviews companies that can disrupt a vertical, don’t build an AI lawyer bot, become a law firm with AI lawyers.’ There is a huge amount of money, compute, water, electricity and resources going toward owning the entire stack of an industry. What does this tell us about where the institutional investments are going vis a vis PropTech? It seems VCs are done funding feature additions; they want companies that replace entire real estate functions, not tools that assist brokers, but systems that make brokers irrelevant. The bar has shifted from “does this improve the workflow” to “does this own the workflow end to end.”
Let’s shift back to Dubai. If I am a brokerage owner, a broker, or a real estate buyer or seller, what I want to know is whether AI will disrupt this space. There are around 40,000 registered brokers in Dubai, over AED 1.5-2 billion worth of transactions per day, and a steady stream of ancillary tech services popping up everywhere to grease the wheels this relatively large industry.
There are market intelligence tools to make decisions more informed (supplemented by DLD transactions). Some services draft tenancy contracts, screen tenants, send contracts automatically, provide renters insurance in case the tenant does a runner or loses their job, convert monthly rental instalments into quarterly or bi-yearly cheques for landlords, and of course video and 3D tools for users to explore homes before visiting them among many other promising startups. While all these disparate tools hold value, as separate entities users have to find them, figure out how to use them and eventually bind them together.
AI and consumer tech today are like Lego blocks. We know how to build ane market these blocks with conventional or vibe coding or co-piloting an app or webApp, and the dream is that they find product-market fit automagically. That happens perhaps 5% of the time where you build it and they will come; most of us are trying to find some semblance of signal in the noise and figure out what people really want and build for that.
The Dubai Land Department made the prudent decision to digitise its sale and rental processes, capturing buyer, seller, landlord, and tenant data at the point of contract signing. With Contract F and tenancy contracts now fully digital, there is a clean audit trail and full transparency for both the government and the public — enabling more informed, data-driven decisions by all parties.
It is important to remember that rentals and sales happen infrequently — rentals once a year, if that, and sales, once in a lifetime. These are not high-volume but high-margin transactions, and buyers and tenants are careful not to act without their armour on. They take their time, reviewing documents, market sentiment, speaking to family and friends, studying real estate data — a fragmented process, yet one that carries enormous weight in the buyer or tenant journey. Property Finder’s 2025 UAE Market Pulse survey found that even among buyers actively browsing listings, the average planning horizon before purchase is six months or more, reaffirming just how deliberate the decision-making process is.1
Brokers usually fill this void, solving problems and expediting decisions as best they can. They are driven purely by profit and operate in a true market economy where buyers and sellers determine the price of properties. Those of us who have been around long enough remember the Wild West between 2005 and 2008 — speculation and flipping were part and parcel of the trade, and we were learning how to row while the canals were still being dredged. Developers building without escrow accounts, buyers issuing fake cheques, agents chasing and threatening sellers from developer offices all the way home by car (this really happened).
While brokers are incentivised to find you what you want, ultimately it will be on their terms, they screen listings, agents, and sellers before a deal can be done. Brokers will kill a deal that doesn’t suit them before they consider whether it suits you. For example, if there is a perfect Springs Type 3M for AED 4.5m, vacant and facing the lake, that fits your requirements perfectly but is exclusive with another agent, that listing — assuming you are fully entrenched with your own agent, will never reach you.
This is the fundamental inefficiency that has persisted throughout the lifecycle of this industry. Exclusivity arrangements are rare, but if even a fraction of brokers operate this way, pocket listings and off-market deals quietly disappear from the buyer’s view. The value a good broker brings is threefold: knowing their market better than anyone else, accumulating listings at fair market value, and marketing those listings effectively. Remove any one leg of that stool, particularly the listings — and the whole thing collapses, leaving brokers retorting to insincere tactics just to close.
Another problem is that brokers have to be multi-modal — they need a functional grasp of real estate law, debt, bank procedures, government services, and tenancy laws, among other things. A good broker will have a reasonable grasp on all of these, but it is impossible to keep up to date with all of this, especially in a market as fast as Dubai where laws and norms change every few weeks.
Brokers are only paid on successful completion of a transfer or rental. Sales are slow, pedantic and often complicated, where 60% are mortgage deals are mortgage deals2 and can take anywhere from 30-60 days to complete. Contract Fs, liability letters, valuation reports, final offer letters, and haggling between buyers and sellers on minute details all delay or potentially kill deals, which are tenuous to begin with. Brokers therefore have to wait for months to receive their commission, of which 40-50% is handed over to their company for the support they provide (which is another kettle of fish to discuss next time).
The system does work, somehow. The DLD has put in place guardrails: brokers now have to pass a real estate exam, get a good conduct certificate from the police, and conduct themselves within RERA’s laws. Remember, there are 40,000 of these guys running around and their main goal is to close the deal before their competitor does and it is dog eat dog; you can put a cage around a savannah, it doesn’t make it a zoo. While competition and the profit motive are drivers of good service and expediency, they also act as an implosion event for the local real estate market. 2025 saw 214,912 real estate transactions which comes out to around 17,900 deals per month.3 This means that just under half the 40,000 brokers in the market are doing direct deals and more likely than not ‘good brokers’ are doing multiple deals, so around 20-25% of brokers are actually walking away with a cheque in hand month end. Rental deals can cover the gap to some extent, but it is not sustainable. Rental commissions are smaller, the 5% fee is split with the other agent and often the brokerage too, and if you factor in the time spent, the numbers rarely add up. There is also a cultural dimension — brokers tend to look down on rentals as easy, low-stakes work, and frankly, they are right that rentals do not keep you sharp in the sales game, which is where the real money is made.
How and where then does AI fit into this workflow? Can it wedge itself into the current paradigm or does it have to disrupt the whole industry to make its point and truly effect change? Buyers want good properties at the fair market price and don’t want to be caught left out of deals that may be right for them. Sellers want to sell their properties at the right price and not leave any money on the table.
Right now, AI is in its nascent stages. As mentioned, we can create the lego blocks to fill this void, through Claude Code, Codex, Lovable, Cursor, and a myriad of other tools. These tools run well locally or within small circles of users, but whether they can be assembled without engineers into a coherent products that remains valuable at scale is yet to be seen. Another question: is there inherent value in transactions between people, and therefore is there an inherent threshold amount you are willing to use AI for versus without — i.e., would you buy a house with only AI, or is it too early for that? In the short- to medium-term, real estate brokerages will remain largely unchanged. As a buyer or a tenant, you will still go to viewings with agents, give them your offer, they will convey that offer to the landlord or seller, and with a finite amount of back and forth, a deal will be reached. However, the profit motive and pure market dynamics hurt all parties in an oversaturated market and I think this is where AI can change the game.
I think in the next 3 years you will be able to spin up apps at the touch of a button and the App Store will eventually become obsolete. You won’t need WhatsApp, Google Maps, or the PropertyFinder app. All the disparate apps mentioned above — from marketplace intelligence tools to rental agreement creators to listing aggregators — will be rolled up and consolidated into one interface. The form factor could be a text or voice app, a virtual agent with an avatar, a pin that listens as you speak and outputs to your phone, or even an app on your glasses or contact lenses. That said, information overload is a real thing, and what humans have is an understanding of the cadence of a deal. Sometimes a client will be busy or overwhelmed, and as humans we recognise that — for real estate, AIs will have to learn it. You will have a digital twin or a personal assistant who will do everything you need and buying a home will simply become where and how you want to live relative to your means. You will no longer have to be paranoid about agents trying to add top-ups on deals, not providing you with the best unit that checks every single one of your boxes (regardless of how trivial or granular they may be).
Where brokers held the reins in deals, you will now have full autonomy. Instead of finding listings on listing platforms, you will speak to your AI agent, tell it exactly what you want and it will find it, whether it is listed or not. It may prospect potential sellers via a message on their AirPods or a message on their IM service that there is a serious buyer interested in your property. If the seller bites, the buyers AI would package their KYC, proof of funds and a formal offer for the seller. The crucial step, the AI would clearly tell the buyer that their offer is over or under the asking price and having studied the market and recent transactions, will give the buyer a comprehensive and tactical guideline on what and how to offer. Assuming the AI has a profile on the seller (their other assets, their profession, their hobbies) which remain private but essential for the bidding process, the AI can guide the buyer to make the offer that makes sense to them.
The AI will then create a chat group with itself as the agent, with the Seller or the landlord. The profit motive has been removed from the broker’s side; the impetus now remains with the main parties to come to an agreement at a fair market value. Given the AI knows the property, the buyer and the seller quite well, it is now objectively the best negotiator for this particular deal.
This is something many startups have tried to do, bring buyers and sellers together — but fundamentally it does not work. When real estate Buyers and Sellers enter a marketplace, they are effectively entering an arena, and you don’t hug and kiss your opponent and walk off into the sunset. It is a form of sparring until you agree on the right price and terms for the deal and this is the real value agents offer, managing egos and expectations.
Furthermore, real estate still works on precedence, which is why good brokers who have experienced the pain of a deal gone pear-shaped can guide you through complicated transactions. Likewise, AI agents would have learnt every deal that has happened since the dawn of real estate in Dubai, and are in fact learning as they process many deals at the same time, thereby approaching perfect information in an industry that never really had it. However, no one likes a know-it-all, and the cadence of the AI agent would be of the utmost importance when designing it, so as not to overwhelm the user.
And therefore, a neutral third party is required: the AI real estate agent. The agent will form a veil between all parties, acting as the mediator, passing along messages, providing guidance, and ultimately closing the deal from start to finish. Information asymmetry has long defined this industry, but with all transaction data available via government sources and no broker acting as the sole arbiter of truth, AI agents can finally close that gap. These AI agents will provide up-to-date, verified data, and their goal is to get deals done while managing expectations. The buyer and seller will not talk to each other unless both turn the key, as you would to launch a missile, and even then they would not exchange contact information unless expressly requested by both parties. The impartial AI agent would remain the negotiator throughout the deal — impartial, truthful and legally compliant from start to finish. The most important thing here would be to really understand emotion and intent of all parties involved. This must be the cornerstone of this paradigm shift — managing expectations, emotions and timelines at a regulated pace that is understandable for all parties.
Everything else falls well into place. Banks, mortgage services, valuation appointments, developer NOCs, Dubai Land Department meetings and cheque transfers all become easier and more transparent for all parties with dealing with an unbiased and neutral AI. There will have to be a ‘human in the loop’ to go out and open doors for viewings and valuation appointments, speak to tenants, go to DLD transfers and pick up cheques from banks but this can be outsourced to a third party conveyance company, sort of like last mile delivery for online marketplaces.
What’s more, as the AI closes more and more deals and recursively self-improves, learning from it’s wins and it’s losses, it will now be able to ‘calibrate’ deals in an optimal way. ‘This is a finance to cash deal with 50% LTV from the buyer and there are 2 sellers which each own 50% equity in the property, one seller is in the country and the other is travelling but seller 1 has power of attorney for the other and the deal needs to be complete in 22 days? We just did something similar, let me bring it up for you and I will adjust the timelines, amounts and banks accordingly.’ No need to explain everything from the bottom up, the AI will already have done dozens of deals similar to this.
This could even see the erosion or complete redundancy of deposit cheques. As all parties now have full knowledge of the partciulars of the deal and there is no longer any information assymetry, the potential of the deal falling apart reduces dramatically. In this case, the 10% deposit could drop to 5% or 1% — deposits still need to insure against job loss, divorce or buyer’s remorse, not just bad faith.
What does this do to the real estate agents in the market? I do not know. Sanitising the reality of AI is a feeble attempt at burying our heads in the sand. I take it all with a pinch of salt: AI is not there yet, not by a long shot. The code these models generate does not scale at the level the founders of the AI labs purport, and adding 1 billion or 1 trillion parameters or context windows will not automagically get us to AGI, and therefore ASI. There are deeper questions about the nature of conscience and sentience to be had by smarter people than me — however, we are not near AGI. Agents, though, I think will for the most part become obsolete. The bigger deals above AED 5m will likely have a human in the loop, but most of the modalities will be handled by the AI; the agent will be a translation layer between it and the client. An important fact to remember: roughly 28% of buyers in Dubai are above the age of 46, and older folks are unlikely to talk to a phone to buy a AED 12m Meadows villa — they would rather have a human. But that human will have to be augmented by AI; otherwise their marginal utility drops to zero.
This is my vision of the future of real estate sales and rentals. AI agents spun up at the whim of buyers, sellers, tenants, and landlords to sell or rent homes. It is within grasp. The building blocks are there. It’s just a matter of putting them together.
Any takers?
3 https://dubaipropertyinsight.com/blog/dubai-real-estate-market-2025-record-transactions
- 1 https://economymiddleeast.com/news/72-percent-of-uae-home-seekers-and-sellers-plan-property-purchase-in-next-six-months-new-survey-reveals/ ↩︎
- https://www.agbi.com/real-estate/2026/01/dubai-property-boom-leans-on-cash-as-mortgage-lending-lags ↩︎
- https://www.agbi.com/real-estate/2026/01/dubai-property-boom-leans-on-cash-as-mortgage-lending-lags ↩︎