AI that delivers real results in your business.

We help mid-sized companies use AI where it reduces manual work, makes information easier to access, and supports better decisions.

We examine your workflows, train your team, and implement the right AI solution, from existing tools to custom software. If AI offers no economic advantage, we say so.

AI Værk × DiZwei

Next event · free of charge

KI-Praxiswerkstatt

AI in your company, but where do you start?

Two hours on site in Süderbrarup: examine one of your own processes, clarify the requirements and leave with a first test step. For owners and managing directors of small and medium-sized companies. The workshop is held in German.

  • 27 August 2026
    6–8 pm

DiZwei, innovation and meeting spaceteam Allee 16–22, 24392 Süderbrarup

Owners and executives in conversation during a workshop

Fewer repeated questions. Knowledge stays available. An internal AI assistant for your company.

We set up an AI solution that answers employees’ questions using information from approved company documents. Every answer names its source and respects access rights. Your team finds reliable information faster, and specialists no longer have to answer the same questions repeatedly.

Explore the knowledge chatbot

Manuals
Operations manual, QM, workplace safety
Policies
Leave, expenses, IT usage
Project documents
Proposals, minutes, specifications

Permissions & processing

Access follows roles: who asks determines the sources
Indexing and linking across all formats
Hosting and processing in Germany or the EU
Question from the team · Production How many days in advance do I need to request special leave, and who do I ask?
Answer with evidence · personalized for Production
You request special leave 14 days in advance directly with your shift supervisor. For your area, the cover rule in chapter 4 also applies: the covering colleague must be named in the request.
Leave policy, section 3.2 Operations manual, chapter 4
1SourcesExisting documents stay where they are. Nothing gets moved around.
2ProcessingRoles and permissions decide which sources are read before every answer.
3QuestionAsk in plain language: in chat, on the intranet, or inside your existing tools.
4AnswerEvery statement carries its evidence. What is not in the sources is never claimed.

An AI solution has to pay off and work in daily operations.

It solves a concrete business problem.

Before the project starts, we define what should work better and how you will recognize the benefit. If there is no useful role for AI, we do not recommend an implementation.

It runs reliably and securely.

Once AI works with company data, interfaces, or important workflows, it needs clear responsibilities, controlled access, and traceable documentation.

The team can use it in daily work.

We involve employees early, test the solution on real tasks, and establish clear rules for use, review, and approval.

We treat your project as if it were our own company.

We assess benefits, effort, and follow-on costs before recommending an investment. More budget is committed only when the next step is justified for your company.

Justus & Elias · Founders · More about us

Entrepreneurial

Only implement what pays off

We weigh the expected benefit against effort, risk, and ongoing cost. If a simpler solution is a better fit or AI offers no economic advantage, we say so plainly.

Treat your budget like our own

We start as small as makes sense, agree on clear decision points, and produce reliable results before committing further budget.

Technical

Choose the simplest viable solution

We configure existing software, connect current systems, or build from scratch. What matters is solving the task reliably, not which technology is easier to sell.

Plan operations from the start

Hosting, maintenance, support, and continued development are not postponed until after launch. We plan them as part of the solution and remain technically responsible.

Case study

Kirsch & Drechsler an AI partnership across the company.

Kirsch & Drechsler has built, sold, rented, and managed residential property in Potsdam since 1992. The company has a team of 15 and has built more than 500 apartments that it continues to manage. Today, AI Værk supports Kirsch & Drechsler as a long-term AI partner across several business areas, from strategy and training to internal applications, customer communication, and joint product development.

Read the case study

A perspective that reaches beyond the project.

We bring technical practice into policy and European perspectives back into our work with companies: our own research and experiments with models, work across the Baltic Sea region, and contributions to European AI infrastructure and AI sovereignty.

More about us

Like this website? We designed and built it ourselves. We can do the same for you. See our web design offer