Module 1
Foundations for automation
Why automation, interoperability and AI need the same stable foundation – and what happens without it.
Lessons
1. Structure before tools
Automation depends on information connecting in a predictable way. A system can only act automatically if it knows which unit an employee belongs to, who the manager is and which cost centre applies.
AI does not fix flawed foundations. An AI service reading contradictory data amplifies the contradictions instead of correcting them.
- Define how information should connect.
- Establish processes that create the information, with one master source per set.
- Set up an organisation that governs the rules over time.
- Adapt the systems that subscribe to the information.
2. Five recognisable problems
When the foundation is missing it shows in daily work: double registrations, hard integrations despite integration tools, manual handling, slow access for new hires and manager data that differs between systems.
The problems look like separate IT issues, but often share a root: the information lacks a shared definition and a clear source.
- Double registrations – the same data is entered in several systems.
- Hard integrations – the technology exists, but concepts differ.
- Manual handling – lists and email fill the gaps.
- Slow access – new hires wait for permissions.
- Inconsistent manager data – systems disagree on who the manager is.
Example
A new hire starts on Monday but gets system access only on Thursday, because the unit has different names in the HR and access systems and someone has to reconcile by hand.
3. Master data and subscription
Master data is the core information many processes and systems need: organisation, employee, location, customer and product. Each information set is created in one process and one system – the master source.
Other systems subscribe to the information instead of registering it themselves. A change then only needs to be made in one place.
Key takeaway: Always ask: where is the information created, and who may change it?
4. Internal and external code lists
External code lists, such as SSYK or SKR's AID, are owned by authorities and organisations and used for reporting. Internal code lists, such as positions, are owned by the organisation itself.
Internal values are mapped to the external ones rather than replaced by them. The organisation can then govern its own structure while still reporting correctly.
Key takeaway: Never invent values in external code lists – use the register's standard.