We stopped selling hours, because we stopped needing them.
A day rate prices our time. You are not buying our time, you are buying a working system, and once AI-assisted delivery cut the time required, billing by the hour would have converted our efficiency into your cost saving only if we chose to pass it on. So we changed what we sell instead.
The commitment
Every proposal states what conventional delivery would cost and take, beside our number. If we cannot beat it meaningfully, we tell you and decline the work.
Three ways to work with us
Pick the one that matches the shape of the problem.
Fixed scope
A defined piece of work with a clear finish line
One price, agreed before we start
The default, and what most engagements should be. We write the scope, price it, and carry the risk of our own estimate. If it takes longer than we said, that is our problem rather than a change request.
Written scope with an explicit definition of done
Fixed price stated against the conventional cost of the same work
Payment on milestones you can verify, not on elapsed time
Change is a new priced increment, never a surprise on an invoice
Retained team
Continuous product work with shifting priorities
Monthly, by team shape
A named team with a fixed monthly cost, working your backlog. Suited to product companies where the roadmap moves and a fixed scope would be re-negotiated every fortnight anyway.
Named engineers, not a rotating pool
Thirty days notice either way, no minimum term
You set priority every sprint, we hold the standard
Monthly written report covering what shipped and what did not
Product partnership
Building something you intend to own and sell
Reduced fee, with terms
For founders building a product rather than commissioning a system. We reduce the fee in exchange for terms that are negotiated in writing at the start. We have declined more of these than we have accepted.
Reduced cash cost against agreed commercial terms
You own the code outright regardless of the arrangement
A defined exit if either side wants out
We say no unless we would use the product ourselves
The arithmetic
Where five times faster at a fifth of the cost actually comes from.
It is not one saving. It is four, and the fourth one is the reason the first three are possible at all.
Code that writes itself firstBoilerplate, migrations, tests and the first draft of most modules. A person reviews and owns every line, which is still far faster than writing it.
~40%of build time
Reading a system nobody understandsLegacy discovery used to be weeks of archaeology. It is now days, and the map is more complete than the one a person would have produced.
~70%of discovery time
Review before reviewOrdinary mistakes are caught before a human looks, so senior review time is spent on design rather than on typos and missing null checks.
~30%of review time
Not building the wrong thingThe largest saving by a distance, and it has nothing to do with AI. A written scope agreed before we start removes the rework that consumes most projects.
The restand the hard part
What we will not do
Four things, stated so you do not have to ask.
Bill for a change we caused
If we underestimated, that is our problem. A change order is for a change you asked for, and the difference is written into the contract rather than argued about later.
Take work we cannot beat the market on
If a standard product fits, we will tell you to buy it. We have talked several clients out of a rebuild, and it costs us the engagement each time.
Hold your code hostage
Repository access from day one, full ownership on delivery, and a documented exit. Payment disputes are handled commercially, never by withholding what you own.
Put a junior on it and call it senior
You are told who is on your team and at what level. There is no pyramid where the people you met are not the people who build it.
Next step
Talk to the engineer who would run your build.
No discovery call with a salesperson, no deck. A senior engineer reads what you send and replies with a real assessment, including when we think you shouldn’t build it.
