promerget.ai
The red flags are always there.
A shareholder resolution that was never minuted. A change-of-control clause nobody has read since 2019. An IP assignment the freelance developer never signed. A customer contract with a most-favored-nation clause that quietly caps the price of every follow-on deal.
None of it is hidden. None of it is hard to find. It is just found late.
Week four
This is how almost every transaction in the Mittelstand — Germany's owner-led mid-sized companies — plays out.
The letter of intent is signed. A price range is set. The data room opens. Three weeks later the other side delivers a findings list: twenty pages, sorted by severity, with the word material far more often than anyone on the sell side expected.
Every item on that list is now a negotiating tool. Not because the buyer is unfair — but because that is exactly what a findings list is for. The price was set against an assumed state of the company. The list describes the gap between assumption and reality. Someone pays for that gap, and at this point in the process it is always the seller.
Why this happens
Not because anyone is careless. Because of when the reading gets paid for.
Reading ten thousand pages of registry extracts, contracts and correspondence was always expensive. Nobody spends that money on suspicion. You spend it when a transaction is running and the cost can be justified against a price.
Which means: the reading happens after the price is set. The economics of the billable hour place the discovery exactly where it does the most damage to the person who built the company.
That is not a failure of due diligence. Due diligence works. It is a failure of sequence.
What changes when reading costs nothing
Once reading a document costs practically nothing, there is no reason left to wait for a transaction that justifies the effort.
CIPHER, our agent for legal review, processes registry filings, commercial contracts and litigation files across DACH, the EU and the US. Half a million documents, 96 percent accuracy, roughly ten times the throughput of a human review team.
The speed is not the achievement. The speed makes the achievement possible.
The achievement is that the finding moves. It happens twelve months before the data room opens — while the IP assignment can still be signed, the missing resolution still passed, and the customer contract still renegotiated in a conversation that has nothing to do with a sale.
When the buyer's findings list then arrives, there is nothing on it that has not long been fixed. The list stops being a pricing tool and becomes what it always should have been: a confirmation.
What this does not solve
An agent makes no legal judgment. It finds, classifies and flags. Whether a clause is a deal risk or a footnote is still decided by a lawyer — and we decide it together with one.
Nor does it repair a badly run company. If the contracts are bad, knowing early just means knowing early. But knowing early is the difference between a problem you solve and a problem someone else prices in.
The uncomfortable part
Most of what gets billed as legal due diligence is reading. The judgment — the part that actually needs a lawyer — is a fraction of the hours.
As long as reading was the expensive part, that model held. It no longer holds, and the firms that adapt will be the ones that bill for judgment instead of volume.
For a seller, the conclusion is simpler and more urgent: the findings list that decides your price is being written right now. By someone. The only question is whether by your side or theirs.
Johann Horch is the founder of promerget.ai. He has led ten acquisitions and integrated seven brands in five countries onto one platform.