AI adoption in companies
Many companies have bought licences for Microsoft 365 Copilot or another AI chat. A few weeks later only a few people still use it. We make sure AI becomes part of your teams’ daily work: with use cases for every role, training on real tasks, clear rules and measured usage.
Microsoft 365 Copilot · Company chats · Local AI
Why AI often fails to take hold in daily work
It is rarely the technology. We come across five reasons again and again.
No link to people's own work
A general training course shows what AI can do. It does not show what the accountant or the service technician will actually use it for tomorrow.
Unclear rules
Nobody knows for sure which data may go into which tool. Out of caution, it is then not used at all, or used secretly with private accounts.
Disappointing first results
Anyone who starts without guidance gets weak answers and does not try a second time.
Nobody to ask
There is no contact person in the team who can help quickly with questions.
Success is not measured
Without figures on usage and time saved, you can neither adjust course nor justify the next licence.
How we support the rollout
For one department or the whole company, in this order.
Collect use cases for each role
In short conversations with the teams, we find the tasks where AI helps in daily work: texts, summaries, research in their own documents, analyses. The result is a list for each role.
Set the rules
An AI policy on a few pages: which tools are allowed, which data may go into them, when results have to be checked and who the contact person is. We involve data protection and the works council early on.
Training on real tasks
Short training sessions for each role with examples from their own daily work instead of generic slides. Each group receives templates for its most frequent tasks.
Key users in every department
One or two people per department who receive more in-depth training, answer questions and pass on good examples. Plus a regular drop-in session.
Measure usage and adjust
We track who uses AI for what and where things get stuck, and adapt training, templates and rules.
What your teams learn
Every role needs something different. That is why we train them separately.
Management and executives
Where AI saves time and money in the company, what risks exist and which decisions must not be delegated.
Employees in the departments
Writing good instructions, checking results, recognising limits and knowing which data must not go in. Practice is based on their own tasks.
IT and administration
Permissions, data access, logging and integration with Microsoft 365, ERP or document management.
Data protection, works council, quality
How the tools handle data, what is logged and how rules can be checked.
Independent of the tool
The approach does not depend on the product. It works with Microsoft 365 Copilot, with a company chat such as Langdock or ChatGPT Enterprise and with a local assistant on the AI Cube. Employees need to know what they use the tool for in their role.
If no tool has been chosen yet, we compare the options with you in terms of data protection, cost, existing software and your teams’ tasks.
When you do not need us for this
If only a few people are to work with AI and clear rules already exist, a single training session is often enough. The formats are on the AI agents and training page. A supported adoption project is worthwhile when several departments are involved or many licences remain unused.
How you can tell it is working
The number of licences says little about the benefit. So before the start and afterwards we measure how many employees use AI at least weekly for their tasks. We add the time per case in the selected processes, review included, and the share of results that need rework. We also record policy breaches and reported incidents.
What the EU AI Act requires
Under Article 4 of the EU AI Act, companies that use AI must take measures to ensure the AI literacy of their employees. The law does not require a specific certificate. Our certificate of attendance lists the topics covered, so you can document the training.
Since August 2026, transparency obligations also apply. A chatbot that talks to customers, for example, must be recognisable as AI. Following the AI Omnibus, the obligations for high-risk systems, for example in applicant selection, apply from December 2027 or August 2028 respectively. We clarify before the project whether your plans are affected and involve legal advice where needed.
If a company introduces an AI tool that can record the behaviour or performance of employees, the works council generally has to be involved. We prepare the necessary information and involve the works council early. This does not replace a legal review in the individual case.
What you have at the end
- A list of use cases for each department, sorted by benefit
- An AI policy agreed with data protection and the works council
- Training materials and templates for the most frequent tasks of each role
- Trained key users in the departments
- KPIs on usage, time saved and quality, before and after
Frequently asked questions about AI adoption
We already have Copilot. Why do we need support?
A licence does not show anyone what the tool is good for in their own work. We find these tasks together with your teams and practise them on real examples.
How do we get employees to actually use AI?
With use cases for every role, training on their own tasks, a short AI policy and key users in every department who answer questions. We measure usage before and after the start.
How do we introduce AI without data protection problems?
With a clear policy on which tools are allowed and which data may go into them. This includes data processing agreements and permissions that follow existing access rights. For highly confidential data, local operation is an option.
Do we need an AI policy?
Yes, as soon as employees use AI tools for their work. It does not have to be long. Everyone should know which tools are allowed, which data must not go into them and when results have to be checked.
Does the works council have to be involved?
If the tool is capable of recording the behaviour or performance of employees, generally yes. That applies to many AI tools because they log usage. We recommend involving the works council from the outset.
What does the EU AI Act mean for us?
Companies that use AI must take measures to ensure the AI literacy of their employees. Transparency obligations have applied to customer-facing chatbots since August 2026, and stricter rules for high-risk applications will follow later. We clarify before the project which obligations affect your plans.
How do we measure whether the adoption was successful?
By regular use in everyday work, the time per case in selected processes, the share of rework and compliance with the policy. We collect these values before the start and afterwards.
Putting unused AI licences to work
Bring along to the initial consultation how many licences you have and how many of them are in use. If you do not know exactly, we clarify that first. The initial consultation is free.
