Workflow — integrations

Two systems, and a human copying between them.

Hours from time tracking into invoicing. Daily takings from a till or web shop into the sales ledger. Documents from a document platform into the right client file. Every time somebody moves data from one system to another, an integration is waiting to be built.

Run the numbers

How often does that transfer happen by hand, per month?

4 times
 

time saved per month

Worked example: ± 90 min per transfer today, zero once the integration runs.

What the workflow does

Reliable matters more than fast

On a fixed rhythm

Every night, every hour, or on every change — whatever suits the process. An integration that runs once a day but always runs is worth more than one that is real-time and occasionally stalls.

With a log

Of every transfer it stays on record what moved and when. If something goes wrong, it can be traced exactly which records made it across and which did not.

Without duplicates

The integration knows what has already been transferred. A restart after a failure therefore does nothing twice, and a record changed on both sides is flagged rather than overwritten.

With an alert on failure

An integration that stops quietly is more dangerous than no integration at all. If something fails, you hear about it — and no message means everything ran.

No API? Often still a way. Where a package allows no integration at all, other routes remain: a scheduled export, an import folder, a CSV dropped at a fixed time, or Peppol. Less elegant, equally reliable — and still better than somebody retyping.

Before and after

A recurring transfer

Today

  1. Open both systems side by side
  2. Export or select the data
  3. Retype or rework the file
  4. Import it and check what came across
  5. Fix duplicates and typing errors
  6. Start again next week

± 90 min per transfer

With the workflow

  1. The integration runs at the agreed moment
  2. The data makes the crossing on its own
  3. Duplicates are excluded by design
  4. You only step in when there is an alert
  5. The log shows exactly what came across

± 0 min

Indicative figures for illustration. An integration is a one-off build with a recurring return: the more often the transfer happens, the faster it pays for itself.

Frequently asked questions

About integrations

What if one of my packages has no API?

Then we look for the next best route: a scheduled export, an import folder, a CSV dropped at a fixed time, or a Peppol connection. Not elegant, but reliable — and still better than a colleague retyping.

What happens if an integration fails?

You get an alert, with what went wrong and which records did not come across. An integration that stops quietly is more dangerous than no integration, so that scenario is explicitly designed out: no message means it ran.

Can data get lost or duplicated?

Every transfer keeps track of what has already been moved, so a restart does nothing twice. If a record differs on both sides, it is flagged rather than one version overwriting the other.

Where does an integration like that run?

On a secure server in Europe, or inside your own environment if you prefer. The data goes no further than the two systems you are connecting.

Free workflow scan

Where does your time actually go?

We walk through your processes together, at your pace. You will know what can be automated in your firm, what it would save, and what is better left as it is. No strings attached.

OSKO Consult is committed to donating 1% of profits to good causes.