What is the Model Context Protocol (MCP)? A clear guide
MCP is the “USB-C port” for artificial intelligence. We explain how the protocol works and why it matters so much for businesses.
AI agents & assistants
AI agents that get real work done
An AI agent is more than a chatbot: it understands goals, uses your tools, checks its results and asks when something is unclear. We build agents that reliably work inside your processes – with clear boundaries and human approval where it matters.
Why it matters
AI agents
Many teams spend hours on repetitive knowledge work: sorting requests, copying data out of documents, preparing quotes, keeping customers informed. This is exactly where AI agents shine. They read unstructured content, decide based on your rules and take action directly in your CRM, ERP, help desk or inbox.
We are model-agnostic: depending on the job we use Claude, GPT, Gemini, Mistral or open-source models running on your own hardware. Through the Model Context Protocol (MCP) the agent gets controlled access to exactly the systems it needs – nothing more.
What you get
Scope of work
Agent design
Goals, tools, roles and guardrails – properly documented before a single line of code is written.
Tool integration
Secure access to CRM, ERP, email, calendars, files and internal APIs via MCP or REST.
Human in the loop
Approval steps for critical actions such as sending emails, payments or contract changes.
Evaluation & tests
Test cases from your daily work measure quality, accuracy and cost before every release.
Monitoring
Logs, cost tracking and alerts – you always see what the agent is doing.
Handover & training
Your team understands and steers the agent itself. No black box.
Typical uses
What companies use it for
- Triage the inbox, classify requests and draft replies
- Create quotes from customer requests, price lists and CRM data
- Research agents for market, competitor or supplier analysis
- Support agents that resolve tickets or hand them to the right team
- Internal assistants for HR, IT and onboarding questions
Process
How we work
Understand the process
We walk through today’s workflow and define measurable goals.
Prototype in days
A first agent works with real (anonymised) examples.
Integrate
We connect your systems and set up permissions, approvals and logging.
Operate & improve
We measure, learn from mistakes and expand step by step.
Technology
What we work with
- Claude
- OpenAI GPT
- Gemini
- Mistral
- Llama (local)
- MCP
- LangGraph
- Python
- Node.js
FAQ
Frequently asked questions
What is the difference between a chatbot and an AI agent?
A chatbot answers questions. An AI agent pursues a goal: it plans steps, uses tools such as your CRM or inbox, checks intermediate results and completes tasks – within clearly defined limits.
How do you stop an agent from making mistakes?
With tight permissions, validation rules, automated tests built from real examples and approval steps for critical actions. On top of that, every action is logged and traceable.
Is our data used to train AI models?
No. We use business APIs and EU hosting options where your data is not used for training. If required, we run models entirely on your own infrastructure.
How long until the first agent is in production?
A prototype is usually possible within one to two weeks. Stable production typically follows after four to eight weeks, depending on the integration effort.
Further reading
Clearly explained: AI, agents & MCP
AI agents vs. RPA: what is the difference – and which suits you?
Rule-based bots or adaptive agents? An honest comparison – and why the best solution is often a combination.
All services
Related services
- 02 · Model Context Protocol
MCP server development
We make your software usable by AI – with custom Model Context Protocol servers.
- 03 · Business automation
AI business automation
Invoices, orders, reports and more run on autopilot – with AI where rules alone fall short.
- 04 · Retrieval-augmented generation
AI knowledge assistants (RAG)
Your company knowledge on demand – with sources, permission checks and answers grounded in your own documents.
Ready for your first automation?
Book a free AI check. In 30 minutes you will know where AI makes the biggest difference in your business.