A pram, just in case
An architecture studio offered to design a house for me because my company delivers on time. A funny email, but it points to something that is not funny in software: work nobody looked at before it went out.
This week an architecture studio emailed me with an offer to design a family house. We have never met or spoken. As far as I can tell, their reasoning went like this: every project on my company's website was delivered on time, so I must care about things happening on time, so I must be ready to build a house.
It made me smile, because by the same reasoning I should also be buying a pram, just in case. But the email stayed with me for a different reason. I doubt a person wrote it, and I am fairly sure nobody read it before it went out. Someone gave a tool a list of companies and let it run. The result was what a tool produces when nobody is watching: well written, with my company's name in the right place, and with no idea who it was talking to.
One day
A reputation is built over years and lost in a day. That is an old line, and it always referred to big things. A failed project, a lost lawsuit, a bad product on the market.
That studio lost theirs in one email. Not because the email is bad. Because it shows that nobody there looks at what leaves the building. If this is how they send offers, how do they draw plans? Who checked the structural calculations? The same hand that checked my website?
Twenty-five years
I have been in this business for twenty-five years. I always approached projects broadly, and that took me into the most demanding phases there are: startups and early stages, when everything is built from nothing. There are no instructions there. You decide what the product is, how people will use it, what its architecture is, what the deadlines are, how it gets made. All at once, and all on you.
Today, with AI, development is no longer the topic. Code is written faster than we can read it. What is left is the person standing behind it, answering for the essence of the product. What we build, why, and what we will not build.
More projects than ever
I talked to colleagues who have rounded off their careers evaluating projects, mostly for investors. They all say the same thing: more projects than ever, more bad projects than ever.
Think about it and it makes sense. A bad project manager or a bad developer usually could not finish something that was bad in its concept. They got stuck. The project died halfway, which was painful, but honest.
Today AI can finish it. It will package it visually, it will look good, and at first everything will look right. It is not. A bad concept is now a bad concept with tidy documentation, clean code and a landing page that promises.
Both feet
This is an enormous risk for companies, and I do not think they realise how far it can push them towards the edge.
Without the new technology you shoot yourself in the foot. Everyone knows that, which is why everyone is in a hurry. With a bad application of it you shoot yourself in both. And immediately, not in a year.
Working with a bad company used to be visible. Deadlines slipped, then came the excuses, then the promises. You had time to react. Now you get the delivery at once. Unconsidered, without a plan, generic, not following and not understanding what you actually need. The project manager sees half of the features for the first time in the meeting with you, and has literally no idea how any part of the system he is selling you works.
You used to pay for slowness. Now you pay for speed, and you get the same thing.
For now
Today this problem mostly shows up as small, misfired messages, like an architect offering me a house because my projects ship on time. It is easy to laugh at those. They cost the sender a little reputation and cost me thirty seconds.
The same pattern is already moving into bigger things. A company that sends emails nobody reads will also deliver software nobody has really looked at. It will arrive quickly, it will look finished, the demo will go well, and there will be no one on their side who can explain how it works or why it was built that way. That is much harder to spot than a strange email, and much more expensive to discover three months later.
So the question worth asking of any partner, tool or agency is not how fast they can deliver. It is who looks at the work before it reaches you, who is allowed to say no, and who answers when it breaks.
I will not be buying the pram. But since that email, I read every offer I get with one question in mind: did anyone on the other side actually look at it?