Over 15 years, the appDev team has worked on more than two hundred projects, from clearly
scoped initiatives at mid-sized companies to large-scale programs at major enterprises. A large
share of that work is mid-market and quick to reach production. That is exactly where our focus
lies. Below is a selection across industries and sizes.
Case studies
Selected projects.
26 projects
AI and RAGsmall to mid-sized
Document recognition with handover to the ERP
Document processing with a human in the loop
Task
Receipts and orders arrive as PDFs, scans and emails and are currently entered by hand.
Solution
Extraction of the relevant fields, a proposal for approval, handover to the ERP. The final sign-off stays with the team.
OCR
Azure OpenAI
.NET
REST to the ERP
PostgreSQL
Fixed-price pilot, then expansion · Fixed-price pilot
Manual prep work drops noticeably, while approval stays with the clerk.
Electronic health record for a large statutory health insurer
Native Android app in the telematics infrastructure
Task
Insured members need to manage documents, permissions and modules of the electronic health record securely on their smartphone, in a strictly controlled environment with strict audit requirements.
Solution
A new interface on an existing white-label base, FHIR data models, sign-in via the sectoral identity provider, consent management, integration of further specialist services such as the e-prescription and emergency data. Complete auditability and documentation designed for the Gematik audit.
Kotlin
Jetpack Compose
Dependency Injection
FHIR
Telematics infrastructure
sekIDP
Jenkins
CI/CD
Team of around ten people, multi-year · Contract-based work within the product team, full code with the client
Passed the Gematik audit and in regular operation.
Internal sales assistant for product data and terms
RAG assistant over product and pricing data
Task
The internal sales team looks up product details, prices and availability across many catalogs and systems, which costs time on every inquiry.
Solution
An assistant answers questions from the approved data sources with source references and respects roles and permissions. Connected to the inventory management system.
Node.js
Vector database
Azure OpenAI
role-based permissions
REST to the inventory system
Fixed-price pilot, then operation · Fixed-price pilot, full code with the client
Answers come with a source, and searching across multiple systems is no longer needed.
iOS and Android with shared logic via Kotlin Multiplatform
Task
An application needs to behave identically on both platforms without maintaining the business logic twice.
Solution
The database and API client written once in Kotlin Multiplatform, embedded in a SwiftUI and a Compose interface. Shared tests, separate native interfaces.
Kotlin Multiplatform
Swift
SwiftUI
Kotlin
Jetpack Compose
GraphQL
Apollo
SQLDelight
Fastlane
Team of around eight people · Contract work within the product team
One shared codebase for the logic, two native apps in operation.
AI integration into an established ERP and legacy landscape
AI features connected to existing systems
Task
A company wants to use AI while keeping its established ERP and legacy landscape.
Solution
Integration with SAP and database interfaces, controlled data handover, AI features layered on top of the existing systems. The existing investment is preserved.
.NET
Node.js
SAP interfaces
REST
Vector database
Docker
several months · Contract work
The AI features run on the existing system, which stays fully intact.
An established application runs on old infrastructure and is hard to maintain; it should become more stable and quicker to roll out.
Solution
A step-by-step move into containers, automated pipelines, monitoring and a clear operating path. Ongoing operation is preserved throughout the transition.
Docker
Kubernetes
Terraform
GitHub Actions
Monitoring
Azure or AWS in the EU
several months · Contract work, modernization and handover
The application runs on modern infrastructure, and releases are automated and traceable.