Questions before a first project – and our answers

An AI employee is an AI agent assigned to a role in your company: with access to your systems, with memory for your rules and with limits at which a human decides. A chatbot answers questions; the AI employee carries out tasks. In detail → AI employees

An AI agent carries out tasks in systems, an AI assistant supports a person in their work, a chatbot only answers questions. The AI employee is an agent with the permissions, the memory and the limits of a role. In detail → AI employees

“Legal duties as a system rule” means that the system can require a mandatory entry instead of just reminding you of it. Examples are the supporting document as a required field, the VAT treatment based on goods and counterparty, and the second approval before a payment. In detail → CRM & ERP with AI

AI consulting maps your workflow and writes down the logic. AI integration implements it in your systems, and in AI operation we keep the result running if you wish. We do all three, always with the process first. In detail → Services

A logic document records your process in your own words: decisions, data sources, exceptions and the points at which a human approves. You sign it off because we only build after that, and that way the result fits your company. In detail → How we work

Working with us suits companies with 5 to 250 employees that use several systems or have a task that no standard software covers. Whether it pays off for your workflow is shown by measuring before and after. In detail → Home

No, this is not only for large companies. The smallest pilot covers one process or one role with one AI employee. A company with five employees is enough for that. In detail → Home

Sometimes industry software is a better fit, and then we tell you so. But if your company works differently from the rest of the industry in three places, the industry software starts holding you back. We check this against your process. In detail → CRM & ERP with AI

We advise against a project if nobody in the company has time to go through the process with us, if the baseline cannot be measured or if your existing system already solves the task. In detail → Services

The price of a pilot depends on the scope of the process, on the systems and on the company. We give you the price range in the first call, based on your process, and in writing within two working days. In detail → Pricing & pilot models

A pilot has a fixed scope and a fixed price. We estimate extensions on a time-and-materials basis before they start. Operation costs a monthly amount that depends on roles, systems and hosting. In detail → Pricing & pilot models

The cost of ongoing operation depends on whether we or your IT operate the system and on how many AI employees and interfaces are running. It is fixed before the pilot. In detail → Pricing & pilot models

For digitalisation and AI projects in SMEs there are federal and state funding programmes. We check in the first call whether your project qualifies. We make no commitment, because the programmes change. In detail → Pricing & pilot models

Memory, agent logic and business data are stored on servers in Germany, separately for each customer, or on your infrastructure. Critical data is processed offline by locally running models; external language models only receive non-critical data, and their providers are processors under Art. 28 GDPR. In detail → AI employees, operating models

Your data is stored on servers in Germany or on your own server, and critical data is processed there offline. Non-critical data can go to external language models whose providers are bound as processors under Art. 28 GDPR. In detail → AI employees, operating models

In the pilot we work with your data in your systems. In the demo, all input counts as test data: real customer or personal data does not belong there. It is deleted when the access expires. In detail → AI transparency

We process critical data offline with local models; it does not leave your environment. Non-critical data can go to external language models whose providers are contractually bound as processors under Art. 28 GDPR. The agent's memory belongs to you. In detail → AI transparency

We recommend what your process and your IT can support: our servers in Germany, separate for each customer, your own infrastructure or a mix. Critical data stays offline in every case, and external models only receive non-critical data. In detail → AI employees, operating models

For a clearly defined process, you see first results 2–6 weeks after signing off the logic document. Visible means measured: time per case and error rate against the baseline, not an impression. In detail → AI integration & process automation

A pilot takes 2–4 weeks while your operations continue. If equipment, sensors or machines are involved, it takes 3–6 months, and before that we measure the baseline for 2–4 weeks. In detail → Pricing & pilot models

The implementation starts with two or three sessions to map the process and 1–2 weeks for rules, permissions and approvals. Setup and integration then take 4–12 weeks, depending on the areas and interfaces. In detail → CRM & ERP with AI

The first call takes 30 minutes, by video or phone, without a presentation: we look at one workflow from your day-to-day work. Within two working days you then receive one page with our assessment. In detail → Contact

In the first 30 days we get to understand your process in one or two sessions and write the logic document, which you read, correct and sign off. Then the build begins. In detail → How we work

Above all, you contribute time: for a CRM or ERP implementation, two or three sessions with the people who do the workflow today, one hour for the logic document, feedback during the pilot and the decision on the switch-over date. In detail → CRM & ERP with AI

Integration usually works. We connect whatever has an interface, a file export or database access: scales via their protocol, phone systems via API, DATEV via document exchange. Anything that has none of these we check in the first call. In detail → AI integration & process automation

The AI needs the data your process already produces today: emails, documents, lists. You can start with Excel. If a workflow depends on a single file, we recommend moving its data into a system with an interface. In detail → AI integration & process automation

You don't need your own IT staff. One person who knows the process is enough: for mapping the process, signing off the logic document and giving feedback. We take over operation, or we onboard this person to do it. In detail → AI integration & process automation

You don't need additional licences for your ERP or CRM; your existing ones remain. Language models and operation are part of our offer, and the cost of operation is fixed before the pilot. In detail → AI employees

We measure whether the AI employee is worthwhile before and after, with your numbers: time per case, queries, errors, onboarding time. We define what is measured before the pilot. In detail → AI employees

Before the pilot we measure the baseline, for equipment the hours of downtime, the unplanned failures and the maintenance costs of the last twelve months. Afterwards we measure the same figures. We don't start a pilot without this baseline. In detail → Predictive maintenance & industrial AI

After the project, either your IT operates the system with our documentation, or we take over operation and maintenance with monitoring and a fixed point of contact. In detail → AI integration & process automation

Onboarding takes place on site, at the workplace, with real cases. Each role also gets an AI employee that answers questions: the new driver asks it, not the boss. In detail → CRM & ERP with AI

The knowledge stays: rules, corrections and decisions sit with the agent for the role, not with the person. Whoever takes over keeps working with it. In detail → AI employees

No, AI does not replace your employees. It takes routine work off their hands, such as looking things up, gathering information and retyping, and puts decisions in front of them. Your employees go from clerks to decision-makers. In detail → AI employees

You can recognise a reputable AI provider by three things: it asks about your process first, not about your budget. It measures before and after. And it says openly what it cannot do. In detail → About us

We are not tied to any vendor, we write down the logic before we build, and we operate what we build. We also work with AI employees ourselves in management and accounting. In detail → About us

An external service provider with a fixed scope is right for one process and one pilot. For operating many processes, it is worthwhile to have your own person, whom we onboard, or to leave operation to us. We document everything so that you do not depend on us. In detail → Services

The EU AI Act applies as soon as you use AI in your company. Since 2 August 2026, AI that communicates with people must identify itself as AI under Art. 50. We check the risk category of your project before the start. In detail → AI transparency

Its role and the approval threshold set its limits: a human decides on money, contracts and staff. Anything beyond the role remains a proposal until a human approves it; every action is logged. In detail → AI employees

We do not name reference customers on the website: our customers' processes are their competitive advantage. We arrange a reference call on request if the customer agrees. In detail → Why we don't name customers