Resident parking permits, end to end without a media break
How a Bavarian city processes more than 100,000 applications per year automatically, without an interface between online form and back-office system.
| 1,000 h saved per year | 4× faster processing | 3.6 mo payback period |
The client
A regulatory affairs department in a major Bavarian city
The Kreisverwaltungsreferat (KVR) is the regulatory and administrative authority of the Bavarian city. It runs central public services, from vehicle registration and driving licences to residents’ affairs and the issue of parking permits. The parking permits team processes more than 100,000 applications per year.
The context
Before this project, the city had already attempted to automate several times. Different service providers had run proof-of-concepts. The processes remained difficult, the results never moved beyond pilot status. Several attempts ended without going into production.
The KVR then brought Lunatec on board, with an explicit mandate that went beyond delivery: build internal automation capability. Lunatec was not just to automate one process; it was to put the team in a position to keep automating in the long term. The first process chosen jointly: the processing of resident parking permit applications.
Parking permits were not a random starting point. The application volume is high, the process is clearly defined, the approval paths are stable. An ideal candidate for a first productive RPA deployment, with measurable impact for residents and case officers alike.
The problem
A media break between two systems
On paper, the process looked modern. Residents could submit their parking permit application online through a web form. In reality, only the intake was digital. Behind it, the work was done by hand.
The online form is not connected to the city’s back-office system for parking permits. There is no interface, no shared database. Instead, the form generates an email containing the application data, which is then forwarded internally. Two case officers read the form data out of the email and copied it manually into the city’s custom-built system for resident parking permits.
What this meant in practice
- Around 4 minutes of pure transfer time per application. Each field copied, pasted and checked one by one.
- Two full-time staff permanently bound to a task with no professional added value.
- No processing outside office hours. Applications submitted over the weekend piled up until Monday.
- Several days of waiting time before the resident received a confirmation email, even though the application had long been on file.
- With more than 100,000 applications a year, case officer capacity was permanently absorbed by data entry alone.
The actual decision-making part, checking eligibility, issuing the permit, handling exceptions, was stuck behind a purely mechanical data-entry task. Specialist capacity was lost to retyping.
The solution
An RPA bot takes over data entry in custom system
Lunatec implemented a UiPath RPA bot that does exactly the work that used to be copy-and-paste. The bot reads the incoming form data from the email, validates the entries, and processes the application in tehir system, using the same interface the case officers use.
The bot can do more in the custom system than just create records. It creates new applications, processes existing cases, and amends data. That covers the three main types of work for the parking permits team:
- New issue of a resident parking permit.
- Vehicle change. Transferring an existing permit to a new vehicle.
- Change of address. Transferring the permit to a new residence within the eligible area.
The scope covers vehicles registered in the city. For these cases, no detailed inspection of the vehicle registration document is required. The logic is rule-based and fully automatable. Applications outside this scope continue to be handled by the team. Some applications are prepared by the bot and then finished off manually, for example where written justifications need to be evaluated, which only a human can assess.
Project shape: 12 person-days, two months, coaching from day one
The actual implementation was carried out in 12 person-days over two months. In parallel, knowledge transfer to the internal team began. A Lunatec consultant coaches the city’s RPA team once a month, on a continuous basis. The goal is not for the city to put a finished bot into operation, but for the internal team to understand the bot, adapt it, and tackle the next process by itself.
This second strand is the decisive difference compared with the earlier proof-of-concepts. The previous attempts ended because, once the service provider left, no one could take ownership. Through monthly coaching, internal capability grows during operations, not afterwards.
The results
The figures come from the productive operation of the bot at the KVR.
| Metric | Before (manual) | After (RPA) |
| Processing time per application | approx. 4 minutes | 30 seconds to 1 minute |
| Turnaround time for residents | several days | within one day |
| Availability | office hours only | 24/7 |
| Automation rate | 0 % | 35 % fully automated |
| Case officer capacity | 2 staff bound to data entry | 1,000 h/year freed up |
| Application volume per year | over 100,000 | over 100,000, no bottleneck |
What the figures mean in practice
Three metrics stand out. First, the factor of four on processing time. What used to take four minutes of manual data entry, the bot does in 30 seconds to a minute. Second, the 1,000 hours per year that the team gets back in case officer capacity. More than 35 per cent of applications run through without human intervention. Third, 24/7 processing. Applications submitted on a Friday evening receive a positive response by Saturday morning. The resident receives the confirmation email within one day, not after a week.
Economics
Annual savings come to around €40,000. With 12 person-days of project effort, the investment pays back quickly: the payback period is 3.6 months. That is a metric that matters in a municipal budget context. The business case closes within the same fiscal year.
Relief for the team
The 1,000 hours freed up per year flow back into case work, but into the parts only humans can do. Complex applications that require careful individual judgement, special cases outside the standard mechanics, cases with follow-up questions to the resident. The work that remains is professionally more demanding than the data-entry task the bot has taken on.
The outlook
The parking permits bot is the first productive RPA process at the city’s KVR that earns the name. It is deliberately conservative in design: on-premises, no AI, with a clear scope. It delivers measurable relief and is comprehensible for the case officers. In the public sector, those two things are the precondition for a second process to follow the first.
Internal capability grows in parallel. Through monthly coaching, the city’s RPA team is building the ability to identify, develop and run further processes itself. Lunatec accompanies that journey as a partner, not as a permanent service provider with exclusive access to the automation.
For the city, this means: one productive process, 1,000 working hours freed up per year, a team that is competent in RPA, and a pipeline of further automation candidates. For residents, it means their application no longer sits on a desk, but is processed, including at weekends, after hours, and during peak load.
| INDUSTRY Public sector / municipal administration | SOLUTION UiPath RPA · integration · 12 PD | PARTNERSHIP Live operation · monthly coaching |