A problem to solve
When the way you work does not fit inside any package.
The work runs on memory, on spreadsheets and on two programs that do not talk to each other. It holds as long as somebody remembers — and what you need to know is known at the end of the month, when it is late to do anything about it.
Tell me what wastes most of your timeHow you notice
- Counting by hand what a photograph would say on its own. Hail dents on bodywork, say: two people in front of the same panel never arrive at the same number, while a model that recognises them from the photo counts the same way every time. The quote stops depending on who makes it.
-
Deciding by eye how much to prepare: either it is left over and thrown away, or it
runs short and the sale is lost.
The history has the answer inside it already, though: how many
are needed tomorrow, for that customer and on that day, is the kind of sum machine
learning does better than a habit.
The number is put forward by the program; the deciding stays with you. - Two programs that do not talk to each other: the same data entered twice. In their place a single way in: the data is written once, and it is the program that duplicates it and sends it where it is needed, to whoever needs it, in the shape they need it in.
- In the evening the numbers add up by approximation, and the difference turns up at the end of the month: too late to work out where it came from. A system that watches the numbers while they form gives warning earlier: the gap that comes back on the same round every time, the stock that will not reach Friday, the customer who has ordered less than usual for three deliveries running.
These are examples, not a list: what I write starts from yours.
The question I start from
The first request is almost always a feature: merge two programs, have an app, get a report
out. At the second question something else comes out, and that is the real work.
The
second question is always this one:
Which number do you take on trust today?
How much you really have in the warehouse, how much was sold along a round, what a job costs you. If the answer starts with “more or less”, that is where the problem is: a number known on trust is not a number known. Everything else — the features, the screens, the reports — comes afterwards, and comes by itself.
How you get from the problem to the system
-
First it gets written, then it gets programmed.
Before the quote you get a document with what will be built, under which rules, and what “done” means for each piece. It is yours even if we do not go on from there.
-
Where it helps, I come and see how you work.
Some things do not get told in a meeting: they get seen. Where the real work happens — on a vehicle, on the shop floor, in the warehouse — half a day is worth more than three meetings in an office.
-
It starts without stopping the company.
Go-live is usually done alongside the old system, where there is one, and I stay close to the people who have to use it every day. What comes up in the first month is not a set of mistakes: it is reality making itself heard, and it gets sorted.
When a ready-made package is enough instead
It is worth saying before an appointment gets made: if what you need is already done by a program on the market, that is the right tool.
Built to order is worth it when the way you work is your advantage and no package follows it all the way, not when it is merely different from how other people do it. The difference between the two is the first thing I look at, and I tell you which it is on the first call.
Tell me which number you take on trust today.
The first contact is free, and it is there to work out whether the problem can be addressed. It is not a sales pitch: if this is not the right fit, I say so straight away.
- Email info@azzertech.it
- Phone +39 351 974 7299