Sabine Krüger Caseworker
Caseworker at a VTS citizen service centre in Hauptstadt.
VTS · Verifiland
A composite drawn from the civic-worker perspective and touchpoint mapping in my research.
20 years in public administration. She is the first real person citizens reach when a digital service fails them.
-
1 Doesn't want to apply online
- Waiting room: people excluded.
- No online appointment: shame, frustration, then a wait.
-
2 Asks about documents
- The AI assistant keeps loading.
- The system doesn't understand. No workaround.
-
3 An exception: the AI gets it wrong
- Case no longer fits.
- Policy is prioritised over people.
Where it breaks
How AI changes their work
- An AI assistant answers citizens' document questions before they reach her.
- She mostly sees the cases the system couldn't handle.
- She has to explain decisions she didn't make and can't see inside.
What gets in the way
New tools rarely do what she needs, so she builds workarounds.
She has no way to correct or override the assistant.
She hears about AI errors from upset citizens.
Fewer face-to-face contacts make struggling citizens invisible.
What helps
AI handles routine questions, which frees her time for complex cases.
Her override reasons become a signal that improves the system.
She becomes the team's early warning for AI errors.
A clear escalation route, so nobody is left without an answer.
Their role path
How to spot outputs that contradict the record, and when to override.
Pausing a message, flagging the case and escalating it through the agreed route.
Uses the override and records why.
Lena, the UX/UI designer, what citizens didn't understand and Amira, the product owner, cases that need a decision