When you invest professionally in property, a valuation isn't just a number — it's a basis you'll later have to defend. Before an investment committee, a credit committee, a lender or a client. That's precisely why thorough valuation and due diligence matter more in professional property investment than in a private home purchase, where a rough figure is often enough. Here's why — and what the difference actually means.
The statistical model's blind spot
Most automated valuations in the Danish market are built on an AVM — an Automated Valuation Model. An AVM is a statistical model that estimates a value from historical transactions and a set of variables. It's fast and cheap, and for a typical owner-occupied home in a liquid market the accuracy is often good enough.
The problem arises when the stakes rise. For an institutional decision, two things are rarely acceptable:
- The uncertainty. A statistical model hits a range, not a point. On commercial and investment properties — where there are fewer comparable transactions — the spread can become wide. An estimate with double-digit percentage uncertainty is hard to commit capital on.
- The black box. Can you explain why the model arrived at the number? Often not in detail. And a valuation you can't account for is hard to defend when someone asks.
For a fund manager or credit officer this isn't an academic objection. It's the difference between a number you can act on and a number you have to get verified manually anyway.
The registers are factual — the model is a guess
Denmark has an unusually rich public data infrastructure for real property. It's factual, not estimated:
- BBR (the Building and Dwelling Register) — area, use, year built, technical details.
- DAR / DAWA — the authoritative address and its geography.
- Tinglysning (the land registry) (the land book, e-TL) — ownership, deeds, charges and easements.
- Statistics Denmark — market benchmarks to put the factual data in context.
A register-based valuation starts in these sources rather than in a model. It doesn't say 'we think the property is 1,240 m²' — it pulls that figure from BBR. It doesn't guess the ownership — it looks it up in the land book. The statistics are laid on top of a factual foundation, not instead of it. The result is a number that can be traced all the way back to an official source.
Due diligence: where the money is really lost
The valuation is only one half. The other is due diligence — and that's often where a transaction goes wrong. Not because the number was wrong, but because something in the transaction wasn't as it was presented.
The classic risk is the gap between the seller's documents and the official registers. A registered easement not mentioned in the sales material. A charge that hasn't been redeemed. An area in the prospectus that doesn't match BBR. Individually small things — together, the kind that change a business case after signing.
Manual due diligence catches most of it — but it scales poorly. Chasing the same information across BBR, the land registry and DAR for every property in a portfolio is time-consuming, and under time pressure that's exactly where something gets overlooked. A systematic register reconciliation does the opposite: it holds the seller's documents up against the registers and flags every discrepancy, so it's handled actively rather than slipping through.
AI can read fast — but the registers must win
AI changes the game by being able to read a whole mountain of documents in minutes. But AI can be wrong — it can misread a figure or 'hallucinate' a plausible-sounding detail. In a professional setting it's unacceptable for such an error to silently overwrite an official record.
So the principle is simple and should be non-negotiable: registers beat AI — always. AI is allowed to read the documents and extract data, but the official registers take precedence. If the two disagree on a field, it isn't decided automatically — it becomes a discrepancy a human must resolve. AI's role is to make you faster, not to decide on your behalf.
Auditability: the valuation that holds up in committee
What ties it all together is traceability. A professional valuation must be able to be verified and defended — perhaps months later, perhaps by someone who wasn't there when the decision was made.
That requires every piece of information to carry its source and its timestamp, and the basis to be frozen as a snapshot at the moment of the decision. Then you can document not just what the number was, but why — and what it was based on. It's the same logic an audit trail provides in accounting: not a compliance formality, but what makes a decision one you can stand behind.
For a private home purchase a quick AVM estimate is often fine. For professional property investment the question is different: Can you defend the number? Register-based valuation and systematic due diligence are the answer, because they build on factual sources, flag discrepancies instead of hiding them, and leave a trail that holds.
How Insight Estates does it
Insight Estates is built on exactly this principle. You describe a property, and the platform pulls DAWA, BBR, the land registry and Statistics Denmark and builds a register-based valuation instantly. AI reads your documents in the background and compares them with the registers, flags discrepancies for you to resolve, and freezes a basis you can document. Registers over AI, traceability all the way — built for decisions that must be defended.