Conpeak
Back to the homepage

Services

Four building blocks for processes that work.

From the first analysis to live operation — every service starts from a concrete business process. Below it, the technologies we work with.

Process analysis and prototyping

We talk directly to the people who run the process every day, identify the biggest bottleneck and quickly build a working prototype. That makes it visible early whether the solution actually helps in daily work.

What's included
  • Conversations with users and stakeholders
  • Analysis of existing workflows
  • Identification of manual bottlenecks
  • A working prototype
  • A clear recommendation for implementation

Custom operational software

We build tailor-made software for processes that standard products don't cover well. The solution follows the real workflow – not a predefined product.

What's included
  • Production and order management
  • Internal administration software
  • B2B portals
  • Machine and inventory management
  • Integration of existing systems

AI workflows and agents

We connect language models with clearly defined tools, company data and controlled approvals. The result: reliable AI workflows that support employees and safely prepare or automate repetitive tasks.

What's included
  • Processing of emails and documents
  • Structured data extraction
  • Search across internal information
  • Preparation of cases
  • Human-in-the-loop approvals
  • Integration with existing systems
  • Evaluation, monitoring and error handling

Continuous development and support

After go-live, we stay responsible. We maintain the solution, react to new requirements and keep developing it together with your company.

What's included
  • Direct support
  • Ongoing improvements
  • New features
  • Monitoring
  • Technical operations
  • A fixed monthly engagement

Technology

Technology that fits the task.

We work with Java, Spring Boot, TypeScript, React, Vue, Flutter, PostgreSQL, Docker, Kubernetes, cloud platforms and both local and cloud-based language models, among others. The choice depends on the process, the existing infrastructure and long-term operations.

Show the full technology stack

Backend

  • Java
  • Spring Boot
  • REST APIs
  • PostgreSQL
  • MongoDB
  • Kafka

Frontend

  • React
  • Vue
  • Angular
  • TypeScript

Mobile

  • Flutter
  • Android
  • Firebase

Cloud and operations

  • Docker
  • Kubernetes
  • AWS
  • CI/CD
  • Monitoring

AI operations and on-premise AI

  • Local LLM inference
  • Open-weight models
  • RAG
  • Hybrid search
  • Reranking
  • Model routing
  • On-premise deployment
  • Containerized deployment
  • Authentication / RBAC
  • Audit logging

AI and agents

  • LLM APIs
  • Structured outputs
  • Tool calling
  • Retrieval
  • Embeddings
  • Vector search
  • Agent workflows
  • Evaluations
  • Observability
  • Human-in-the-loop