Sakan positions itself as an "AI-native real estate brokerage" for Dubai's residential market. The thesis: traditional brokerages are people-heavy, low-margin operations where AI can dramatically improve broker productivity (from 3-4 deals/month to 9-10) while reducing overhead.
| Name | Role | Background | Notes |
|---|---|---|---|
| Zade Al Borshaid | Co-founder / CEO | AUB (American University of Beirut), Antler (Dubai, ~2 years running cohorts 6-7), Brevan Howard (summer analyst 2016) | UPDATED Primary experience is Antler VC accelerator, not Brevan Howard. Has seen hundreds of startups from the investor side. Left Antler ~6-7 months ago to build Sakan. |
| Nicolas Bochenko | Co-founder / CTO | Property Finder (Dubai, joined 2012): scaled 4 eng/$20M to 100+ eng/$800M. Then Delivery Hero (Saudi acquisition). Then Amina venture fund CTO/CPO. | CONFIRMED Strong technical co-founder. Built and scaled the region's largest property portal. Deep domain + engineering expertise. |
Dubai real estate is experiencing a cyclical peak, with record transaction volumes and values in 2024-2025. However, significant supply is coming online:
Sakan's one-pager claims a $6 billion commission pool in Dubai. Dubai Land Department (DLD) data shows AED 13.7 billion in total broker commissions for 2025, which converts to approximately $3.7 billion USD.
The $6B figure is either: (a) inflated by ~62%, (b) includes the broader UAE/GCC, not just Dubai, or (c) projects forward to a future year without labeling it as such. Any of these explanations raise credibility concerns.
Ask them directly: "Your one-pager says $6B commission pool. DLD data shows AED 13.7B which is $3.7B. Can you reconcile?"
Pre-meeting research was incorrect. Nicolas confirmed during the call that he joined Property Finder (Dubai) in 2012, scaling from 4 engineers/$20M valuation to 100+ engineers/$800M valuation. He then moved to Delivery Hero for a Saudi acquisition rebuild (scaled to 70-80 engineers), and most recently was CTO/CPO at Amina venture fund.
This is a significantly stronger technical co-founder profile than initially assessed. He has direct experience building and scaling the region's largest property portal — exactly the domain Sakan operates in.
Zade's one-pager lists Brevan Howard as part of his professional background, which implies substantive experience at a top-tier macro hedge fund. Research indicates this was a summer analyst internship in 2016.
A summer internship is not the same as a full-time role. Not disqualifying, but the presentation creates a misleading impression of seniority.
Addressed during the call. Laura (surname not disclosed) is joining as Head of Brokerage in the coming weeks. Her profile:
Her initial role is not managing 200 agents — it is finding and onboarding 1-5 top-decile brokers for the first 18 months and leveraging her 20-year network to open doors. This is the right approach for an early-stage brokerage.
| Company | Vertical | Last Raise | Valuation | Investors | Relevance |
|---|---|---|---|---|---|
| Harper Insurance | Insurance brokerage | $47M Series A (2025) | ~$200M (est.) | Emergence Capital, YC | HIGH Identical thesis: AI-native service agency replacing traditional model |
| EliseAI | Real estate ops AI | $250M Series E (2025) | $2.2B | Sapphire, Point72 | MEDIUM AI for real estate but multifamily ops, not brokerage |
| Compass | Tech-enabled brokerage | IPO 2021 | ~$2.5B (current) | Public | MEDIUM Cautionary tale — tech brokerage, still unprofitable, down 80% from peak |
| Side | White-label brokerage | $150M Series D (2022) | $2.5B (peak) | Coatue, Tiger | MEDIUM Platform model for agents; raised at peak, now in correction |
Compass (US, IPO 2021) is the most relevant precedent for a tech-enabled brokerage:
Quick-reference definitions. Use these naturally in conversation — they signal you understand the tech, not just the pitch.
"Your AI needs data to be smart, but you need to be smart to get data."
AI models learn from historical data. A new company has no historical data. So their AI is essentially guessing until they accumulate enough transactions. Netflix had this — it cannot recommend movies to a new user with no watch history. Sakan has the same problem: their AI cannot match leads intelligently until they have hundreds or thousands of completed transactions to learn from.
Drop it as: "How are you handling the cold start problem? Your lead matching needs transaction data you don't have yet."
"More users → more data → better AI → more users. A self-reinforcing loop."
The best AI companies create a cycle where using the product generates data that makes the product better, which attracts more users. Google Search is the classic example. For Sakan: each closed deal should generate data (buyer preferences, broker behavior, pricing outcomes) that makes the next deal faster to close. If they are not capturing and feeding back this data, they are just a tech-enabled agency, not AI-native.
Drop it as: "What does your data flywheel look like? How does each transaction make the next one better?"
"Built ON AI from day one, not AI bolted on later."
AI-enabled: A traditional business that adds AI features. Like a brokerage that adds a chatbot. The business works without the AI.
AI-native: The business fundamentally cannot operate without AI. The AI IS the product, not a feature. Sakan claims to be AI-native — but if you remove the AI, is there still a brokerage? If yes, it is AI-enabled, not AI-native.
Drop it as: "If we turned off the AI tomorrow, could your brokers still close deals? If yes, you are AI-enabled, not AI-native."
"What stops a bigger player from copying you in 3 months?"
Warren Buffett's concept. A moat is the thing that protects your business from competitors. In tech: proprietary data, network effects, switching costs, regulatory approvals. For Sakan: if they are using off-the-shelf AI (ChatGPT/Claude APIs), Property Finder could add the same features quickly. Their moat needs to be something harder to replicate — like government API integrations or proprietary transaction data.
Drop it as: "What is your moat? Property Finder has the data and the distribution. What do you have that they cannot replicate?"
"How much money do you make (or lose) on each broker?"
The math of one unit of your business. For Sakan: Revenue per broker (their cut of commissions) minus cost per broker (tech, support, compliance, office space). If a single broker is not profitable, adding more brokers just means losing money faster. This is how Compass burned through $1.5B.
Drop it as: "Walk me through the unit economics per broker. What is the fully loaded cost to support one broker, and what commission revenue does that generate?"
"How much does it cost you to sign up one new broker?"
Marketing, recruitment, onboarding, training costs to get one broker onto the platform. If CAC is high but broker churn is also high (common in Dubai — most new brokers quit within 12 months), you are constantly spending to replace brokers who leave.
Drop it as: "What is your broker CAC, and what is your annual broker churn rate?"
"The big AI brain they are building on top of — ChatGPT, Claude, Gemini."
Large Language Models are the general-purpose AI systems from OpenAI, Anthropic, Google. They are powerful but generic. Building on them via API is fast but not defensible — anyone can call the same API. The question is what proprietary layer Sakan adds on top.
Drop it as: "Which foundation models are you using, and what is your proprietary layer on top?"
Dubai real estate regulatory alphabet soup.
RERA: Real Estate Regulatory Authority — licenses brokers and agencies, sets rules.
DLD: Dubai Land Department — registers property transfers, maintains ownership records.
Oqood: System for registering off-plan property sales (before the building is completed).
Ejari: Mandatory rental contract registration system.
NOC: No Objection Certificate — developer must issue this before a property can be resold.
Drop it as: "Are you integrated with Oqood and DLD systems, or is the paperwork still manual?"
"Total market, serviceable market, what you can actually capture."
TAM: Total Addressable Market — the entire Dubai commission pool ($3.7B, not $6B).
SAM: Serviceable Addressable Market — the portion Sakan could realistically serve (e.g., residential only, specific price ranges).
SOM: Serviceable Obtainable Market — what they can realistically capture in 3-5 years.
Drop it as: "Your TAM is $3.7B based on DLD data. What is your SAM and SOM for the first 3 years?"
Every PropTech pitch deck in 2026 says "AI-powered lead matching, automated follow-ups, smart CRM." That is table stakes, not a moat. The word "native" implies AI is the foundation, not a layer on top. That distinction matters enormously.
Standard approach: Buyer says "3BR in Marina, $500K budget" → filter listings → show matches. Any database does this.
The real opportunity: Dubai brokers waste 60-70% of their time on unqualified leads and mismatched showings. A genuinely AI-native approach would need to:
60% of Dubai transactions are off-plan. Commissions are 2-8% (vs 2% secondary). This is where brokers make real money.
But off-plan selling is fundamentally relationship-driven:
The unsexy but defensible play is not lead matching — it is transaction operations:
Sakan's pitch: AI takes brokers from 3-4 deals/month to 9-10. That is a 2.5-3x improvement. Let's stress-test:
| Broker Activity | % of Time | Can AI Help? | Realistic Time Saved |
|---|---|---|---|
| Lead qualification & outreach | ~30% | Yes — filtering, scoring, auto-comms | 15-20% |
| Property viewings | ~25% | No — buyers want to see properties physically | 0% |
| Negotiation & closing | ~20% | No — deeply human, cultural, relationship-dependent | 0% |
| Admin & paperwork | ~25% | Yes — document prep, portal submissions | 15-20% |
If their AI stack is built on ChatGPT/Claude APIs (which it almost certainly is at pre-seed), what stops Property Finder — with 15+ years of listing data, developer relationships, and millions of user sessions — from adding the same AI layer in 3 months?
The only defensible moats for Sakan would be:
If building this from scratch, the architecture should be layered — each layer generates the data the next layer needs:
Document generation, government portal integration, compliance checks, timeline management. This is the "aspirin" that solves today's pain. It also generates the transaction data needed for Layer 2.
Which brokers convert which buyer profiles. Optimal showing sequences. Pricing strategy recommendations based on comparable analysis. This requires transaction data from Layer 1.
Only possible after enough transaction data exists. This is where the "AI-native" promise actually delivers — but it is a Year 2-3 feature, not a launch feature.
Meeting held via Zoom on May 6, 2026
Attendees: Zade Al Borshaid, Katy Huang
Zade spent 2+ years at Antler (Dubai), running 6-7 cohorts as a venture builder. He left approximately 6-7 months ago to validate his own idea in PropTech, driven by a belief that if he didn't start now, AI advancement in the next 3-5 years might close the window. He reconnected with Nicolas Bochenko, whom he had met during Antler's first Dubai residency 2.5 years ago. Nicolas had been trying to find the right co-founder opportunity since then.
Nicolas joined Property Finder (Dubai) in 2012, scaling the team from 4 engineers / $20M valuation to 100+ engineers / $800M valuation. He was then headhunted by Delivery Hero to rebuild the tech stack for their Saudi acquisition, scaling to 70-80 engineers. Most recently served as CTO/CPO at Amina venture fund. He is building the entire Sakan tech stack alone using AI tools (Gemini, Claude, Cursor), with no plans to hire additional engineers for the first 9-12 months.
Zade's core insight: existing brokerages are operationally resistant to AI transformation. He cited the Air / Unique Properties JV failure in Dubai — an attempt to retrofit AI onto a large existing brokerage that collapsed because the brokerage refused to change its operations. His thesis: you must build a brokerage from the ground up with AI-native operations, not bolt AI onto legacy processes. Dubai's market is extremely fragmented — 95% of brokerages have fewer than 10 employees.
The first months were spent validating with brokers, brokerages, developers, and advisors. Current build priorities:
Laura is joining in the coming weeks as Head of Brokerage. Her profile: 20 years in Dubai RE, trained 5,000+ brokers, currently C-suite at a top firm, prior PropTech experience. Her initial role: find and onboard 1-5 top-decile brokers for the first 18 months, leveraging her 20-year network. She has been advising Sakan and validating all workflows.
Launching in approximately one month. Strategy: give free AI tooling to brokers to generate conversation data for the flywheel. Zade: "The more conversations we have, the more data we generate, the quicker — that gives us an edge in terms of the flywheel."
Zade was candid: tech is a 1-2 year moat at best. The real defensibility comes from:
Dubai first (9-12 months) → Abu Dhabi → Saudi Arabia (100% local broker ownership required) → Southeast Asia (Malaysia, Vietnam, Thailand, Indonesia). European interest: A $30B+ European residential/commercial RE investor with 60,000 units is interested in piloting Sakan's tooling for their build-to-lease portfolio — potentially bringing 80% of externally managed lettings in-house.
Currently finalizing their pre-seed raise. Terms, valuation, and amount were not disclosed during the call.
| Item | Pre-Meeting Assessment | Actual (from call) |
|---|---|---|
| Nicolas / Property Finder | WRONG Flagged as PropertyGuru (Singapore) | CONFIRMED Property Finder (Dubai). Joined 2012. Scaled 4 eng/$20M → 100+ eng/$800M. Then Delivery Hero, Amina Fund CTO. |
| Zade / Brevan Howard | FLAGGED as overstated | CONTEXT Primary experience is 2+ years at Antler (Dubai), running 6-7 cohorts. Brevan Howard was minor. Antler experience is actually more relevant — he has seen hundreds of startups from the investor side. |
| Brokerage experience gap | FLAGGED Neither founder has ops experience | ADDRESSED Laura joining as Head of Brokerage. 20 years Dubai RE, 5000+ brokers trained, C-suite at top firm. |
| $6B commission pool | FLAGGED DLD data shows $3.7B | NOT DISCUSSED Was not raised during the call. Still unresolved. |
Katy asked: "How can you fight the likes of Anthropic, OpenAI — especially if you are building on top of them?"
This is the single most important strategic question for any AI startup. It forced Zade to articulate his real moat thesis (relationships + brand + market share, not tech). His answer was honest and thoughtful — he acknowledged tech is a 1-2 year moat at best. This question alone demonstrated serious investor-grade thinking.
Katy suggested: "Patent the bottle, keep the recipe."
This is genuinely valuable advice for an AI startup. You cannot patent an LLM prompt chain, but you can protect the orchestration architecture, the data schemas, and the process IP through trade secrets while patenting the customer-facing interfaces. Zade clearly had not considered this — he was stuck in a binary "patent everything or patent nothing" mindset. This reframe could be strategically significant for them.
Katy advised: "While you are building for Dubai, leave it open so you can work in other solutions — regulatory requirements differ by market."
Practical and correct. Dubai (RERA), Saudi (100% broker-owned requirement), and Southeast Asia (local ownership rules) all have different regulatory frameworks. Building a rigid Dubai-only system now creates expensive rewrites later. This is the kind of advice that saves a startup 6 months of refactoring down the road.
Katy suggested white-labeling the tooling for larger companies. This is actually validated by the conversation — the European RE investor ($30B+, 60K units) is essentially asking for a white-label deployment. Zade didn't connect these dots explicitly, but the European interest proves Katy's instinct was right. A dual revenue model (own brokerage + white-label SaaS) could accelerate monetization.
Katy correctly positioned MAKR's stage preferences: "We prefer Series A and above — validated product, revenues on the books." This was honest, set expectations appropriately, and didn't shut the door. Zade understood: "Good for us to think about as we look over the next 12 to 18 months." If Sakan executes, they could be a MAKR pipeline deal at Series A.
Katy casually mentioned that Tarin produced a 3-part Anthropic analysis overnight using deployed AI agents. This was perfectly placed — it demonstrated (a) MAKR uses AI at an advanced level, (b) Katy understands orchestration architectures firsthand, and (c) she is not a tourist asking surface questions. Zade's response ("Awesome, I'd love to hear") showed it landed. Credibility established without showing off.
The thesis is sound: existing brokerages will not transform themselves (Air/Unique JV proves this), so you must build from scratch. An AI-native brokerage that handles lead qualification, nurturing, and transaction ops end-to-end is genuinely differentiated in MENA.
But "better mousetrap" requires two things they do not yet have:
Without these two proof points, they are just a small brokerage with a tech story. With them, they are genuinely disruptive.
| Criteria | Assessment |
|---|---|
| Stage fit for MAKR | TOO EARLY Pre-seed, no product, no revenue. MAKR targets Series A+. |
| Sector fit | ALIGNED PropTech / AI — fits Smart Living vertical. |
| Team quality | STRONG Nicolas (Property Finder scale-up) + Laura (20yr brokerage) + Zade (Antler VC experience). Complementary skills. |
| Market opportunity | LARGE $3.7B commission pool, genuine whitespace for AI-native brokerage in MENA. |
| Thesis quality | SOUND Build-from-scratch approach validated by Air/Unique JV failure. Right thinking on moats. |
| Current investability | NOT YET No product, no revenue, no unit economics. Revisit at seed/Series A with proven metrics. |
Sources consulted for this analysis. Use for further review or independent verification.
Note: This analysis was produced on May 6, 2026 using a combination of pre-meeting research (AI-assisted, multi-source) and post-meeting transcript review. Market data reflects conditions at time of writing. All figures should be independently verified for investment decisions.
Need to understand where they are in their fundraise and at what terms.
Are they generating revenue? How many brokers are on the platform? How many transactions have they facilitated?
Clarify Nicolas's actual background. The one-pager says "Property Finder" but research shows PropertyGuru.
Many AI startups pitch future vision. What is live, tested, and producing measurable results right now?
Revenue per broker, cost to support per broker, CAC for acquiring a broker, churn rate. This is the entire thesis.