111,000
historical tickets, searchable for the first time
Customer story · Public sector
Answer faster, search less.
How komuna GmbH, after years of searching, found a partner that delivers what it promises — cutting search times, easing the load on support, and making 111,000 historical tickets accessible in seconds for the first time.
- komuna GmbH · Altdorf bei Landshut
- Software for municipal administrations
- 80 employees
- In production since July 2026
Starting point
200 requests a day. And no shared knowledge base.
Software provider komuna supports around 700 municipal administrations in Bavaria and Baden-Württemberg — with 80 employees and software whose development and support demand deep procedural expertise. The support team answers around 200 requests a day on average, and up to 400 at peak times.
Before Daeo OS, answers rested almost entirely on the experience held by individual employees. Today, ticket history, FAQs, training materials, and manuals are connected into a searchable institutional memory — GDPR-compliant and source-cited.
Avg. 200
support requests per day, peaks up to 400
~3 months
from data audit to productive use
4
specialist teams on one shared knowledge base
Before and with Daeo OS.
Before
With Daeo OS
Before
Knowledge in individual heads — lost with every departure
With Daeo OS
A central, source-cited institutional memory — it stays in the system
Before
111,000 historical tickets, but unfindable — full-text search returned only raw data
With Daeo OS
All 111,000 historical tickets usable as a source — comparable cases and earlier solutions found in seconds
Before
Each specialist team with its own FAQs and lists — little exchange between teams
With Daeo OS
One shared knowledge base for all four specialist teams
Before
Every request costs search time — employees often had to ask colleagues
With Daeo OS
Search times substantially shorter, requests handled faster
Before
During illness or holidays, whoever covers starts from scratch
With Daeo OS
Seamless handover — including the customer's history
Before
Data only as current as the last manual update
With Daeo OS
The system syncs every five minutes — knowledge stays current
Before
No measured support metrics — nobody knew how long a request took or how often it was resolved on first contact
With Daeo OS
Monthly insights into request topics and knowledge gaps
Before
Onboarding new colleagues: six to nine months
With Daeo OS
Goal: working independently within days
First real test with the hiring wave from September 2026
Results
What actually changed for komuna.
01
Noticeably higher productivity
Substantially shorter search times and faster resolution of support requests. Employees report “a huge time saver”, mainly because comparable earlier cases can now be found in seconds.
02
Historical knowledge finally usable
What used to be unfiltered raw data is now solution knowledge that can be retrieved on demand — with a source cited for every answer.
03
Cover without a break
When someone is ill or on holiday, whoever covers picks up seamlessly from that customer's earlier requests instead of starting from zero.
04
Knowledge stays in the company
The experience of departing employees is documented and retrievable, instead of leaving the company with them.
05
Support becomes measurable
A monthly insights module shows the most frequent request topics and knowledge gaps — the data foundation that was previously missing entirely.
06
Goal: onboarding in days, not months
New colleagues previously needed six to nine months. komuna expects a substantial reduction with Daeo OS — the first real test runs with the hiring wave from September 2026.
It was never a question for us that AI might replace a job here — quite the opposite.
Why the old setup no longer held up.
komuna's software can determine whether a budget gets approved or an election can be run — so the support team needs deep procedural and product knowledge, especially at peak times such as year-end or during legislative changes like the property tax reform.
Yet the answers were collected nowhere: 111,000 historical tickets sat in the database without any layer of abstraction — full-text search returned only unfiltered raw data. In practice, almost everything was answered from the experience of individual employees, and each team had built its own FAQs and lists that were rarely shared between teams.
The consequences were costly: long onboarding for new colleagues, duplicated work when two people solved the same case independently, and the repeated loss of valuable knowledge whenever experienced employees left the company.
The decision
Look at the data first. Then make a promise.
An earlier attempt with an off-the-shelf chatbot had failed. In its search for a suitable solution, komuna had heard plenty of big promises from vendors — but they rarely held up once things got specific. deAI Labs took the opposite approach and analysed komuna's data before making any claims about the expected outcome. It was precisely this methodical approach that convinced them.
Data ownership and GDPR compliance were also decisive for management and the IT department — as was the fact that the knowledge employees retrieve through Daeo OS updates continuously from the real data.
What you promised, you delivered — and that still holds today. You keep every commitment you make, and it simply works.
We do not need a standard chatbot — for that I could just use ChatGPT. We need solutions built for us that can be extended and expanded over time.
The rollout
From data audit to production in about three months.
March 2026
On-site data audit
The first meeting took place before any proposal: where do requests come from, how are solutions searched for, where is time lost? Only then came the technology.
Connection
Sources joined into one memory
Support tickets, FAQ and instruction collections, training materials, and manuals were connected through a semantic abstraction layer — far more than a raw data index.
Before go-live
Test catalogue with agreed quality
komuna and deAI Labs jointly defined a test catalogue with contractually agreed answer quality before the system went into operation.
Since July 2026
In production in daily support
They started where the pressure was greatest: with the ticket crew in support. A monthly insights module makes the results visible as a dashboard with KPIs.
In daily work, employees ask their question in chat in their own words. Daeo OS searches the knowledge base and answers directly in the conversation, with a source cited — and syncs every five minutes with the connected sources. New tickets, updated instructions, and fresh solutions become part of the knowledge base almost immediately.
Data ownership preserved, GDPR-compliant.
Data processing
Covered by a data processing agreement between komuna and deAI Labs.
Built and operated in Germany
EU data residency and full control over your own data.
Traceability
Every answer is backed by its source, rather than a black box.
Contractually agreed quality
A test catalogue before go-live with contractually agreed answer quality.
Support was only the beginning.
Further departments are set to follow step by step. That the collaboration goes beyond support is already clear: by their own account, komuna is “already thinking about completely different projects” with deAI Labs.
All figures are as reported by komuna GmbH, as of August 2026. Quotes translated from German.
Secure your team's hard-won knowledge.
As komuna shows, capturing knowledge in support does not have to be a year-long project. Let's talk through your specific case in a no-obligation first conversation.
Book a first conversation