The XAI Assistant is aimed at people and companies who want to add explainable AI components to their artificial intelligence (AI) application in order to better understand AI decisions. Developers receive targeted information and tips for successful XAI use for a specific XAI use case.

The XAI Wizard provides information and advice on a specific XAI use case to help companies create suitable design alternatives for their XUI. Relevant information is determined on the basis of a questionnaire and output together with examples of corresponding XAI methods. However, the XAI assistant does not provide any direct solution approaches or specific recommendations for XAI methods, as this is not reliably possible based on the current state of research. Instead, the XAI canvas enables interested individuals and companies to obtain targeted information about relevant properties of declarations and provides ideas on how the initial design alternatives can be implemented for further evaluation and iterative further development. Developers are thus given an overview of important aspects and potentially suitable XAI methods.

To use the XAI wizard, the context of use and the usage requirements of an XAI use case should be defined, as this information must be entered in the form of a short questionnaire. The XAI canvas provides assistance in determining the context of use and user requirements.

The XAI wizard first queries important aspects of the context of use and the usage requirements. Based on this information, users receive information and advice on suitable properties of declarations as well as examples of corresponding XAI methods. The XAI wizard was designed together with the XAI canvas and is intended for use on individual XAI user stories.

The questionnaire is structured as follows:

1. Users:
This section asks for information about the end users under consideration.

- Prior knowledge:
Either domain experts or laypersons are considered. Domain experts have expert knowledge in the domain under consideration, but not necessarily AI expertise. An example of this would be an insurance agent. Laypersons have neither AI expertise nor domain knowledge. An example of this would be a person to be insured.
- Need:
A collection of questions from which a question can be selected to be answered by an explanation of the AI decision. This represents the end user's need for an explanation.
- Application:
Additional information about the application context. Internal refers to an application within the company that uses the AI system. External refers to a private end user. This is usually a customer of the company using AI.

2. Explanatory properties:
Three typical properties of explanations can be selected if they are particularly relevant for the specific XAI use case: the correctness of the XAI method used, the comprehensibility of the explanations provided and the time-criticality of the calculation and provision of the explanations.

After completing the questionnaire, the relevant information is collected and displayed in an output field.