Intelligence
AI agents and automations that measurably save your team hours.
We put frontier AI where your team loses the most hours. We build agents and automations on your data, under EU rules and independent of any single model. We modernize old software step by step, without stopping operations. Beforehand we set how many hours the solution should save. Our Security team reviews every system before it goes live.
Where it usually gets stuck
- Lots of AI tests, none of them in operation
- There were workshops, prototypes and a list of ideas. Day to day, the team still works as before. No test made it into a process that someone is responsible for.
- Your data is scattered
- Quotes in the CRM, contracts on a drive, knowledge in emails. A model can only work with what it finds. Without access and clear permissions, every AI stays a chat window without context.
- Old software slows everything down
- A system from earlier years keeps operations running. It is slow, expensive to maintain, and hardly anyone still knows it. A rebuild all at once seems too risky, so it stays as it is.
What you get
Agents that complete tasks on their own: sorting inquiries, preparing quotes, reconciling data. They work with fixed permissions and ask a person before doing anything irreversible.
We connect CRM, ERP, drive and email so recurring work runs by itself. Your team reviews the results instead of retyping data.
We build in line with GDPR and the EU AI Act. We document data flows, permissions and responsible persons in writing. Hosting and data processing stay in the EU wherever possible.
Your solution does not depend on one provider. When a better model arrives, we swap it in without rebuilding the process.
We renew old systems step by step. New parts take over tasks from the old ones while operations keep running. There is no big cutover day and no downtime.
We write down instructions, examples and review criteria. That way your team can carry the solution forward, even when someone else takes over the work.
What stays the same in every project
- Cause first
- Problems rarely arise in one area alone, but where areas touch. We keep looking until it is clear where it comes from. Fixing a symptom is cheaper, but it comes back.
- Technology involved from the start
- We plan tracking, page structure and interfaces while the strategy takes shape. That way you do not find out only during implementation what is not possible.
- You hear the why too
- For every proposal you learn what it is based on and which path we rejected. That way you can take part in the decision and explain it to others.
A good fit if
- You want to use AI in operations, not just try it out.
- Your team loses hours to repetitive work.
- Old software slows your business down, and a rebuild all at once is too risky.
- Your data has to stay in the EU.
Not a good fit if
- We implement AI in practice. For legal advice on the EU AI Act or GDPR, you need a law firm.
- We want to take work off your team's plate. A project that only aims to cut jobs does not fit us.
What you probably want to know
That depends on the processes and systems. After the analysis you get a quote for the first process, with a goal and review dates. We only expand once the first step saves its hours.
The right one for the task. We build model-independent, so you can switch models later. Where needed, we use models with data processing in the EU.
Your data does not train any model. We use business accounts with training turned off. Which data an agent may see, we define together beforehand.
Our Security team. Before launch, it tests every AI system for prompt injection, data leakage and misuse of tools. We map the findings to NIS2, GDPR and the EU AI Act.
Yes, that is what we recommend. For example quote creation or sorting incoming inquiries. The bigger goal stays in view, so the first step becomes the foundation for further ones.
Related offers
- AI agentsAgents prepare quotes, answer inquiries and reconcile data. Your team approves.
- AI automationAI takes over recurring work like retyping, sorting and forwarding. Your team reviews.
- AI chatbotCorrect answers from your own documents, even at night. Difficult cases go to your team.
- Software modernizationNew parts take over the tasks of your legacy system step by step. No cutover day, no downtime.
- AI agency SwitzerlandAI consulting and implementation from one team for Swiss companies. Data protection under nDSG settled before launch.
- AI red teamingWe attack your chatbot or AI agent before anyone else does. Every finding with priority and fix.
- Shopify integrationsConnect ERP, PIM, CRM and accounting to Shopify, with clear data ownership.
- Headless CMSWhen a headless CMS makes sense, which one fits and what it costs.
- Next.jsWebsites and web apps with Next.js. Fast to load, easy to find, in your own code.
- B2B websitesWebsites for offers that need explaining. For purchasing, engineering and management.
6 min read
New AI model? Delete everything first, then set up again
More context used to help a lot. With the 2026 models that flips. Why we delete rules, skills and hooks at every model change and bring back only what is missing.
5 min read
AI news October 2026: what companies should know now
New models from Anthropic and Google, a standard for agents in retail and the first conflicts with agents on the open web. What matters for your company.
5 min read
Agentjacking: how a fake error message takes over AI agents
Three attacks from 2026 show the same pattern: an agent reads data from outside and executes it as a command. What happened and which guardrails help.