Select, establish, and further develop AI

In the 16th and final episode of our AI series on the use of AI in the skilled trades, we complete the implementation process of the digital assistant KaI.

This article is part of the special topic Artificial Intelligence in Crafts – an assistant called KaI

Overall six steps are necessary to use KaI – your digital assistant – effectively in your company:

Step 1 | Observe | Analyze the current state

Step 2 | Identifying the need for action | Developing a strategy

Step 3 | Define the challenge | Select suitable solutions

You might also be interested in:

Operations management

Work more safely and healthily with Kal

Operations management

Kai in everyday work – How do you find your Kai?

The Chambers of Crafts in Germany - Chamber of Crafts and Trades of Saarland

Effective use of AI in the skilled trades

Step 4 | Identify the solution | Test the solutions and processes in pilot projects

Step 5 | Implementation | Implementation and use throughout the entire company

Step 6 | Process Management | Scaling and Further Development

In this episode, we will look at the last three steps.

Implementation of KaI

When making a selection, a distinction can be made between solutions that they develop themselves – perhaps even with the help of AI – and those that are to be purchased. We will assume in the following that they... want to purchase a solution, so that in the fourth step we can combine the selected AI solutions into one test in a small, manageable settingThe goal is not to immediately convert the entire operation, but to determine whether the tool is feasible. From the various solutions available, the next step is to find the one that is suitable for implementation in your company. ID runs in two steps .

Two-stage process

Become first All selected tools were evaluated and tested., and in the second stage, the selected tool is used in a small team or at a Pilot project – depending on the type of implementation idea – extensively testedThey therefore examine the various solutions for their user-friendliness, practicality, and above all, with regard to the established criteria and their new processes. Because every type of digitization or the use of a new tool leads to a change in the process. To ensure the Processes zu to verify, these will be tested in the second test phase complete with In AI system testedAll interfaces and data are also tested, and adjustments are made if necessary.

put a Evaluation matrix to identify the most suitable solution for them. Ingredients: The evaluation matrix should include not only the goals or expectations for the new system, but also the schedule, Budget, technical and legal framework and the Data availability At the end of this phase, a well-founded decision is made: If the solution proves successful, it can be implemented throughout the entire company. If not, improvements are made or an alternative is chosen.

Plan the implementation carefully

If the pilot phase was successful and the solution is to be implemented company-wide, we are now at step five. The phase of Implementation It needs to be planned carefully. This isn't just about the technical introduction, but also about organizational changesSystems need to be set up, interfaces created, and employees trained.

Requirements, which should be given special attention:

• Are all necessary devices, such as smartphones, tablets, PCs, sensors, cameras, etc., available?
• Are the existing systems compatible?
• Is there sufficient storage space and server capacity?
• Do the interfaces between the AI ​​tool and existing programs work?
• Was data security, such as access rights, role models, etc., checked? 

The important thing is that Gradual introduction zu togetherCreate a feasible timelineInstead of changing all work areas at once, a clearly defined process must be implemented in stages. Sufficient time must be allocated to clarify questions, adapt processes, and gather experience. It is equally important to... to offer support – for example, through short training sessions or simple instructions that explain how to use the new solution. Successful implementation is characterized by the fact that the solution is not perceived as an additional task, but rather integrates seamlessly into daily work routines. Employees should realize: AI helps us, reduces our workload, creates clarity, and improves quality.

AI series and application examples All parts of the AI ​​series can be found in our special issue "Artificial Intelligence in Crafts – an assistant called KaI".

In addition, there are further articles about AI on handwerksblatt.de, for example the online articles "A mighty clever way to communicate" about the voicemail app with AI chatbot from the start-up "meiti", "AI – how craftsmen benefit in their everyday business"Or"ChatGPT: Save time and money with AI" from the theme special "Digital Crafts". 

It is also important that all Employees to this Preparations are being made for changeAnd this so-called change management doesn't start now, but with the very first step: analyzing the current situation. The better understood they feel, the smoother the change process will be.

Humans are creatures of habit, and even when the path is difficult, we tend to cling to old—albeit ineffective—routines. This inevitably leads us to question why we're making the change in the first place. However, with thorough preparation, we can mitigate the challenges of the change process and reach the summit more quickly. During implementation, positive results should be made visible, such as time savings or fewer errors. Nevertheless, the fundamental pattern of a change process remains consistent, and it's crucial to prepare employees for this as well.

Active process management

After the introduction, the phase of continuous evaluationAI solutions are constantly evolving – and so is business. Therefore, a active process managementThis includes, regularly zu ., if she Solution the weiterhin the desired benefit brings and where Improvements are possible. During this phase, the following will also be discussed: Scaling Consideration: Can the solution be used in other areas? Can additional functions be activated? Are there new tools that might be suitable for the operation? 

Many craft businesses find that new ideas emerge from their initial experiences. It is equally important to... Focus on data quality zu keepAI systems function better the better the data they receive. Therefore, standards should be developed: How is data collected? Who is responsible? How are potential errors corrected?

Process management also means, in exchange zu stay – internally with the Employees and externally with Suppliers or partner companies. Regular feedback sessions To help identify potential and solve challenges early on. Lifelong learning is therefore not only relevant for individuals, but also for KaI. Thus, a digitalization project is not a project with a fixed end. It is an ongoing process. Nevertheless, if you actively pursue this path, you will benefit in the long term – through more efficient processes, better quality, and more time for what matters most.

The author was an employee of the Mittelstand-Digital Zentrum Handwerk (SME Digital Center for Crafts), a project funded by the Federal Ministry for Economic Affairs and Energy. From 2016 to June 2026, this center supported craft businesses and craft organizations in leveraging the opportunities offered by digital technologies, processes, and business models – free of charge, vendor-neutral, and nationwide. Since 2024, there has been a particular focus on artificial intelligence.

DHB now also digital!Simply click here and register for the digital German Crafts Journal (DHB)!