Real-estate intelligence
Qompas is the independent intelligence layer for real estate — a provenance-honest index that scores what a market or unit is actually worth, calibrated on real transactions, that no one can pay to move. A reference the whole market can point to.
Geometric across four pillars — a single weak pillar can't be bought back by a strong one.
Asking price is as listed. Yield, score and the transacted benchmark derive from DLD transactions and Ejari contracts — the model never invents numbers.
Why we're certain
We tested model after model around AI brokers and sales automation, and hit the same wall every time: buying property is still an act of trust, and automating the sale doesn't close that gap. But running those experiments put us in front of hundreds of real buyers — and showed us what does close it: trustworthy data. We pivoted to the layer beneath the transaction, with two things a fresh team can't buy: first-hand behavioural evidence from real buyers, the conviction of having tested the alternatives to destruction, and real demand for proven global data from every contact we had in the industry.
The problem, and the answer
The product
Built by three engineers and a designer, before this round. Three things prove what a slide can't:
Four-pillar decomposition with live signals that decay on a separate clock from the structural score — news moves the short-term read without whipsawing the number.
Proves: the model is real, not a diagram.Ask for a unit, a comparison, a portfolio — it answers in structured cards drawn from the data, and refuses when it can't ground the answer. It will not invent a number.
Proves: disciplined where every other AI tool hallucinates.Dubai calibrated against DLD, Cyprus shown at mixed confidence, Thailand returning an honest “no calibrated data yet.”. No number is invented.
Proves: the honesty is enforced in the system, not the marketing.How a client actually uses it
Layla advises a family office in Dubai. A client has handed her about AED 40M to invest in property. She is looking at a bundle of off-plan units across Marina, Business Bay and JVC — each showing 7%+ projected returns, with a market story suggesting prices climb next quarter. Nothing in the offer is dishonest; it is simply all quoted from asking prices, and she has no independent way to check them.
A dozen units, each with projected returns — all figured off the asking price. The arithmetic is honest; the reference point is the problem. She has no independent way to tell whether those prices reflect the market.
Qompas puts each asking price next to what units there have actually sold for. Several Marina units are priced ~15% above the real market — so the 7% is honest math on a price that’s too high.
That “prices about to jump” story shows up as a short-term blip that fades over months — kept separate from the area’s score, which hasn’t moved. It doesn’t even claim to know if the news helps or hurts prices.
Area score: unchanged
She passes on the overpriced units, steers the money toward ones priced in line with reality, and spreads it across areas. Her client gets an honest read on the whole plan — and isn’t rushed into overpaying on a headline.
Qompas didn't pick the investments. It showed her what the units are really worth, and which “now or never” deadlines are real — across a large, multi-property plan. That's what she pays for, every time.
Who it's for, and why they pay
Investors and family offices placing capital — and the advisors, brokerages and agencies who now win by letting independent numbers do the talking. First revenue comes from the professional, not the casual browser — and both sides of the table are professionals.
At the decision moment: check the area’s real fundamentals, see asking vs. transacted, and put an independent read on the table — whether you’re the one buying, or the one being asked “how do I know that’s fair?” Repeated every deal.
A few thousand a year is rounding error against a single $1–3M decision — and for a professional, an independent number you didn’t author is the fastest way to be believed. The same logic that makes 355,000 professionals expense Bloomberg.
| The alternatives | Independent | Buy-side decision layer | Transacted, not asking | Instant & usable |
|---|---|---|---|---|
| A relationship & a recommendation | — | — | ✓ | ✓ |
| Portals (asking prices) | — | — | — | ✓ |
| DLD wrappers (e.g. DXB Interact) | ✓ | — | ✓ | — |
| Valuation platforms (e.g. Property Monitor) | — | — | ✓ | — |
| An AI chatbot | ✓ | — | — | ✓ |
| Qompas | ✓ | ✓ | ✓ | ✓ |
Access & pricing
See the full Index map across every location, ask the index for any area, browse every listing, and preview what the Portfolio does. The honest read is open; paying unlocks doing the work at scale.
For the serious solo investor, vetting one decision at a time.
For advisors, agencies and family offices working a book of deals over time.
Enterprise — bulk seats, integration, and custom data access for institutions and family offices. Pricing by conversation.
Market
The category is proven — we enter through Dubai, the most investor-driven property market on earth, and compound from there.
CoStar revenue (2025)
a data-only RE business, +19% YoYDubai transactions, 2025
270k deals · ~193k active investorsof homes sold off-plan
bought with no transaction historyserviceable revenue, Dubai alone
~30k seats × ~$3.6k — before any other marketBottom-up, not top-down. The $250B is market vitality, not our market — data spend isn’t deal value. Our SAM is built from seats × price (CoStar’s floor), at under 20% penetration of the active investor base — and that base widens as the sell-side goes data-driven and agencies buy independent reads too. Then it compounds two ways: more investment-grade markets as each is calibrated, and licensing the verified data layer to AI real-estate products that can’t source neutral data.
Why now & why it holds
Where we are, honestly
The round
on a SAFE · 12 months
Three milestones, each closing one of the three real risks. Hit them and the seed valuation is earned, not asked for.
Convert the developer channel into a verified, calibrated buy-side data feed — and stand up the first full-time core team.
Complete the index backtest against Dubai Land Department transaction history.
Sign the first paying buy-side design-partners.
Use of funds: data acquisition & engineering first; then key-account management, dev/design extension, licensing & infrastructure. Long-term: the verified data layer becomes infrastructure other AI products license.
Team
CEO — four years investigating the real estate data landscape
CPO — eight years leading international products
CTO — a decade in architecting complex digital systems
Plus 2 full-time engineers and a lead designer — the team that built the live product before this round. Two years inside the problem: capital-efficient, and willing to follow the evidence even when it means changing course.
No slide argues as well as a product that refuses to invent a number.
Calibrated · DLD