Any FinTech founder raising in the UK this year will have seen the numbers – according to KPMG’s Pulse of FinTech, investment fell to €2.1 billion (£1.8 billion) in the first half of 2026, down around two-thirds on the same period last year and the lowest since the report began in 2016. Deal count hit a ten-year low. The capital that is still moving is going to AI, and within AI it is going to where the results can be seen fastest.
Alongside the numbers runs a story: AI has collapsed the sales cycle. Products that used to take a year to sell now close much more quickly, and founders who can’t match the pace should question whether they are building the right thing, or building it in the right country.
For anyone selling into regulated financial services, the story is a little different. Worse, believing the first version of the story is one of the fastest ways to put a good company out of business. Rather than being an issue, the long sales cycle is one of the few things in this market that is a strength and cannot be copied in other industries.
What a sale actually looks like
Here is what selling AI to a mortgage network, a wealth manager or an advice firm actually involves. It means integrations into systems that were not built with an innovative AI provider in mind. It means lengthy security reviews, several boards and committees, each on its own calendar. It takes several internal sponsors, who need to be won over separately and who do not all agree with each other. You also need trust in the system, trust in the output, and the certifications to prove it, from the usual ISOs through to the new AI equivalent, ISO/IEC 42001.
In the UK at least, the procurement cycle has not sped up. Regulated financial services firms are accountable to the FCA for what their systems decide, a point the regulator’s own Mills Review made quite plainly this summer, and in compliance much of the work is straight-through: a file review or a financial promotions check is either right or it is a regulatory problem. A highly regulated product, with many integrations, that has to pass several boards and several sponsors, moves as slowly as it always has. The sales cycle is long, and that is fine and normal.
Founders arriving with consumer expectations miss this. They hear about the fastest sales cycle in AI without noticing that the companies telling that story are B2C, or B2B businesses that behave like B2C. The lesson travels but the conditions do not.
Surprise is what kills a startup, not the cycle
A long cycle only becomes dangerous when it is a surprise. If a founder expects deals to close in weeks and sizes their future round, their hiring plan and their investor conversations around that timeline, the business can find itself in real trouble by month nine, with nothing closed and no fault in the product. If the cycle is a known quantity from day one, it is simply a parameter to design around.
Designing around it means a few unglamorous things. Size rounds so the money lasts through more than one full cycle. Hire people who have sold into regulated institutions before and know that a fourth meeting with the risk committee is normal rather than a warning sign. Set investor expectations before the term sheet, not after the first missed quarter – and get the certifications done early, because the security review will ask for them either way.
Our investors describe the length of our sales cycle as a moat, and they are right. The entrants it protects against are the naive ones, and in this market they are usually people from inside the industry – they have very good product knowledge, a real understanding of the problem yet no experience of how running a startup works. They build something credible, start selling, and run out of money before the first contract lands. The technology can be copied and the domain knowledge is widely held. Being set up to wait is the hardest thing to copy, because it has to be built into the funding, the hiring and the expectations from the beginning.
Build in the slow lane, and build it here
The other conclusion founders are drawing from the funding numbers is that conditions will be better in the US and while for some companies it is, for others it would be wise to question what they are actually aiming to build before making that move. The most compliance in the world is in Europe, not America: the most demanding regulators, the most complex rulebooks, and the customers with the strongest reason to buy. Leaving the market where your problem is most prominent, because the deals there take longer, is a strange trade. The US may not be the answer for everyone.
The slow lane has three things going for it. There are fewer competitors, because most cannot afford to wait and many of those that do try, fail. The customers you win stay, because a firm that has just spent a year taking you through three committees has no appetite to repeat the exercise with someone else. And the market rewards the one thing that is hard to fake: a company designed, from its first round, to survive the wait.
FinTech is not dying. Parts of it are, but companies using AI to do work that firms could not do before are growing quickly but selling slowly. The nine-month deal is not a flaw in regulated FinTech. It is the entry fee, and the founders still standing in two years will have paid this on purpose.
