Boiman Kupermann Solutions is a German AI consulting firm with 20+ years of engineering experience. We build production automation systems, run them in the cloud, and train teams to operate independently.
What sets us apart: we run our own business on 49 custom AI skills. From incident handling to invoice tracking, from news generation to meeting transcription — everything is AI-powered. When we build automation for clients, we're sharing what we use ourselves every day.
Production operators processing invoices across the European Peppol network — inbound, outbound, and invoice validation for multi-country enterprises (DE, AT, BE, FR, LU).
Peppol AS4 UBL 2.1 CII XRechnung Storecove EN 16931 SAP ByD
Autonomous agent networks for document classification, intelligent routing, and end-to-end workflow automation. Our AI Labor demonstrates real-time agent orchestration with MCP and A2A protocols.
MCP Protocol A2A Protocol Claude AI Azure OpenAI LiteLLM
Test automation with Playwright and browser-use, CI/CD pipelines, and full observability through the ELK Stack (Elasticsearch, Kibana) and Grafana. We set up the infrastructure that keeps production systems reliable and transparent.
Playwright browser-use pytest GitHub Actions ELK Stack Grafana
End-to-end workflows connecting enterprise systems — email classification, SAP integration, document processing, and Microsoft Graph API automation.
Azure Functions Azure Container Apps SAP Microsoft Graph Infrastructure as Code
This is what makes BKS-Lab different. We don't just consult on AI — we run our entire company on custom AI workflows.
We've built a library of specialized AI skills that automate our daily operations:
| Category | Skills | Examples |
|---|---|---|
| Operations & Monitoring | 7 | Log analysis, deployment verification, incident handling, post-deploy checks |
| Customer & Communication | 6 | Email management (MS Graph), SharePoint sync, Outlook processing, project notifications |
| Development & Projects | 5 | GitHub Projects automation, commit analysis, dashboards, sequential thinking |
| Content & Documentation | 7 | Multi-agent news generator, PDF export, certificates, meeting transcription, wiki analysis |
| Cloud & Infrastructure | 5 | Azure provisioning, blob storage portals, website deployment, backup automation |
| Scheduling & AI Meta | 4 | Recurring task scheduler, context orchestrator, Telegram bot, product vision builder |
Meeting recording drops into folder
→ AI transcribes and classifies meeting type
→ Structured protocol written to wiki
→ Action items extracted
→ GitHub issues created with correct project fields
→ Status notification sent via email
Stepstone reports an invoice error
→ Incident skill queries Elasticsearch logs
→ Correlation ID and root cause identified
→ Incident documented in wiki
→ GitHub issue created with all required fields
→ Go-live tracking updated automatically
One command loads your entire project context — repo paths, open tasks, relevant documentation, active skills, and session state. No manual context-switching.
/cockpit load stepstone # Loads everything for Stepstone project
/cockpit load docemos # Switches to Docemos context
/dashboard --all # All projects at a glanceWe teach what we practice. Our workshops are built from real experience operating AI-native workflows at scale.
| Program | Format | Audience |
|---|---|---|
| Claude Code Mastery | 30 min - 2 days | Developers learning AI-assisted development, multi-agent workflows, custom skills |
| AI Competency Training | Half-day | Entire organizations building AI literacy and practical use cases |
| AI for Leadership | 2-3 hours | Executives evaluating strategic AI adoption and roadmap planning |
| Business Development with AI | Full-day | Product teams using Business Model Canvas for AI product ideation |
All programs include certification according to Article 4 of the EU AI Act — delivered to 29+ professionals across engineering and technical documentation sectors.
graph LR
A["Assess"] --> B["Build"] --> C["Deploy"] --> D["Validate"] --> E["Handover"]
| Phase | What happens | Deliverable |
|---|---|---|
| Assess | Process analysis, data flow mapping, automation potential | Business case & roadmap |
| Build | Rapid prototyping with real data, iterative with your team | Working proof of concept |
| Deploy | Production on Azure with monitoring, alerting, CI/CD | Live system with SLA |
| Validate | Real workload testing, performance benchmarking | Verification reports |
| Handover | Documentation, team training, knowledge transfer | Self-sufficient operations |
| Domain | Technologies |
|---|---|
| Languages | Python, TypeScript, Go |
| AI & LLM | Claude (Anthropic), Azure OpenAI, MCP Protocol, A2A Protocol, LiteLLM |
| Cloud | Azure Functions, Container Apps, Key Vault, App Insights, Blob Storage |
| E-Invoicing | Peppol Directory, SMP/SML, Storecove API, UBL/CII/XRechnung |
| Web | Astro, React, Tailwind CSS |
| Data & Observability | ELK Stack (Elasticsearch, Kibana), Grafana, PostgreSQL |
| Quality | Playwright, browser-use, pytest, GitHub Actions |
| Infrastructure | Terraform, Docker, SOPS, uv, Semantic Release |
graph TD
A["Wiki — The Office"] --> E["Shared Packages"]
B["bks-web — The Showcase"] --> E
C["bks-codex — The Lab"] --> E
D["bks-infra — The Engine Room"] --> E
E --> F["23+ Repositories"]
| Pillar | Purpose | Contains |
|---|---|---|
| Wiki | Documentation & operations | Protocols, customer docs, processes, standards |
| bks-web | Public presence | Marketing website, blog, AI labor demos |
| bks-codex | AI tooling | 49 skills, MCP servers, commands, packages |
| bks-infra | Infrastructure | Terraform, secrets (SOPS+YubiKey), forensics, monitoring |
