Our conviction · Essay
The Truth About Numbers
Markets run on trust in numbers, and the way those numbers get verified has barely changed in a generation. Why we think that is about to end, and why we built a firm to end it.
The numbers we decided to believe
I have been the person wiring the money. I have bought companies and I have sold one for $285m, and in every one of those rooms the most important documents were financial statements that no one present had verified. We priced businesses, moved capital and changed lives on numbers we had simply decided to believe.
Markets are made of trust. Strip the trust out and nothing else functions: not price, not credit, not a single transaction. An investor wires $200m to acquire a company they have never operated, on the word of people they have never met.
That trust does not appear on its own. A profession manufactures it, and then everyone forgets it is there. Accountants close the books, auditors attest to them, tax advisors turn them into obligations and diligence providers decide, before a deal, whether any of it can be believed. They are the trust infrastructure of the entire economy, and almost no one thinks about them until the moment the trust fails.
Here is the thing almost no one will say plainly.
The profession that certifies reality has, across most of its work, quietly stopped checking it.
The language of certainty
Tested5%
Certainty belongs to a claim that can be traced back to its source and confirmed. What the profession too often sells is the feeling of certainty instead: its language, applied to numbers no one has actually verified.
Sample 5% of the transactions, and call the result an audit opinion.
Close the books to a figure that ties to last month rather than to source, and call it a financial statement.
Adjust an EBITDA that three parties calculated three different ways, and call it diligence.
A financial statement is a map, and the profession long ago stopped walking the territory to see whether the map still holds. In good times the gap between the map and the ground stays invisible. In bad times it is the whole story.
Every financial collapse is
an epistemic collapse first.
It is the moment when the shared picture of reality drifts from the facts and it turns out no one had been watching. The savings and loan crisis, the mortgage securities, Enron. Each time the response was to add rules on top of the ritual, never to ask whether the ritual still produced knowledge.
We have reinvented how people communicate, how goods cross the planet and how disease is diagnosed. We have not reinvented how a company's reality becomes a number a stranger can trust.
Why the ritual survives
Everyone inside the system is doing what the system pays for. The legacy accounting firm sells human time, leveraged over a pyramid of partners, seniors and juniors, and it lives on the spread. In that arrangement technology is a threat by construction. Anything that reduces the hours reduces the revenue.
The legacy firm is aligned to hours, so it cannot want to be faster. The software vendor is aligned to consumption, so it cannot want to be leaner. Neither is aligned to the only thing the client has ever wanted: a number that is true, produced as fast and as cheaply as the truth allows.
What the machine makes possible
The work divides into three kinds of labor, and only one of them requires a person.
For a century the profession arranged this the other way around, because the pyramid was the product. Armies of juniors did the reading and the arithmetic while a partner reviewed a sliver and signed. Rebuild the firm around the machine and the pyramid turns over. The system reads everything and ties out everything. The senior professional does the judgment, and supervises, calibrates and stands behind the software that does the rest.
Get the division of labor right, and speed and accuracy stop being a tradeoff. They become the same thing.
What emerges is a firm. It delivers the work, owns the number and answers for it.
Software hands the accountant a tool. A firm hands the client an answer, and stands behind it.
Someone has to verify the machines
Verification is one problem wearing many costumes. The engine that reads an entire general ledger to price an acquisition is the same engine that lets an audit start from everything a company shares rather than a sample. The numbers a company is bought on, audited on and taxed on should finally speak the same language.
There is a harder reason this becomes inevitable. The books themselves are automating. When your own AI writes your numbers, you cannot ask that same AI whether they are true. The institution built to reconstruct financial reality from source is the natural check on automated finance.
The window
If you underwrite your business on manual, sampled, six-week analysis while the person across the table underwrites theirs on the full picture, traced to source, in days, you will lose to them inside a few years. The advantage of knowing what is actually true compounds on every decision, every deal and every audit.
The window is open. It does not stay open long.
A better relationship with the truth
The most important financial technology is the ability to know whether a number is true. Everything else in finance is downstream of that, and almost everything sold as innovation skips it: a faster way to move money, a prettier way to look at it, one more interface laid over data no one has checked.
Institutions do not accept a machine's number because a demonstration impressed them. They accept it when every figure can be traced to its source and a person of judgment has put their name to it.
Trust is earned one verifiable number at a time.
That is why our answer puts the machine and the person together. It starts at the deal, and the same discipline carries into the audit, the close and the tax return. We believe the production of trustworthy numbers will be rebuilt from the foundation within the decade. We intend to be the ones who do it.
Konstantin Mehl
CEO & Co-Founder, Daneel