Your tools don’t talk to each other
So a person is the integration. Every order, every customer, every stock figure passes through somebody’s keyboard twice.
What this looks like day to day
An order arrives in one place and gets entered again somewhere else. Stock is right in one system and wrong in another. The same customer exists three times under slightly different names. Somebody exports a file on Friday so that somebody else can import it on Monday.
This is among the most immediate improvements available, because the data already exists and only needs to move on its own. The moment it does, every downstream number agrees and stays that way.
Wherever you’re starting from
Nothing in place yet
Nothing is connected and a person moves the data. The first integration is usually the largest single win available on this list, and the one whose effect is visible within days.
Something, but it isn’t good enough
Something connects them and you do not trust it. It syncs one way when you need two, drops records quietly, runs nightly when the business needs it now, or was built by someone who has left. A connection nobody trusts gets checked by hand, which means you are paying for the integration and doing the work anyway.
Working well, could go further
Your systems talk and the data holds up. What is left is usually scope: a third system still outside the loop, a field that never made it into the mapping, or reporting that would be possible now that the data finally agrees.
Where this usually ends up
Nothing here is decided in advance. These are the directions this problem usually goes. Which one fits you comes out of looking at how your business actually works.
A real connection
Your systems get joined directly: orders, customers, stock, statuses and documents moving between them automatically, in the direction that makes sense for each.
One source of truth per thing
For every piece of data, one system owns it and the rest follow. Without that decision, an integration just spreads the disagreement faster.
A clean-up first
Duplicate customers and mismatched records get found and merged before anything is connected, because joining two systems is what makes existing mess visible.
Failures that announce themselves
A sync that hits trouble retries itself, and if it still needs attention somebody is told immediately. That visibility is what lets a team trust the numbers without checking them by hand.
How we work out which of those it is
We look at how the work happens now: the actual sequence rather than the process document. What gets done, by whom, how often, and where it waits. That is a conversation and then some time with the people doing it, and it is where the answer comes from. The call and the audit are done by Bruno himself, so nothing you explain has to be explained again to whoever ends up building it.
What comes back is the shortlist worth building and what each piece returns, ranked so the strongest goes first. You see that in writing, with the price attached, before anything starts.
What changes
The re-typing stops, and with it the class of mistakes that came from it. Numbers agree across systems because only one of them is allowed to decide. The work that used to happen between Friday and Monday happens in seconds, continuously.
DUK
Our systems are old. Do they even have an API?
Often yes, and where there genuinely is none there is nearly always another route: a database, a scheduled file, or driving the interface itself. Age is rarely the obstacle it looks like.
Do we have to replace anything?
Usually not. The tools are normally fine; the gap between them is the problem. Replacing a working system is expensive and rarely necessary.
What about the duplicate records we already have?
They get dealt with as part of the work, with an explicit rule for what counts as the same record and who is allowed to create new ones afterwards.
How long does this take?
Two systems and a clear set of fields is typically a few weeks. More systems, or ones that disagree about the same data, take longer, and you will know which of those you have before committing.
Which two systems don’t talk?
Tell us what gets typed twice. We can usually say within one conversation whether it can be connected and what that involves.
Susisiekti → What we solve →