1. Introduction to AI Assistant

Mendix offers MAIA (Mendix AI Assistant), a suite of intelligent tools that assist developers by speeding up development, reducing errors, and improving the user experience.

The goal of these features is to provide intelligent support to developers, whether they are experts or novices, enabling them to optimize their work and accelerate the deployment of applications.

2. MAIA: The Mendix AI Assistant

MAIA is at the core of AI-assisted development in Mendix. It operates across three main areas: guidance, assistance, and component generation.

2.1. Guidance: Interactive Assistance for Developers

MAIA includes an integrated chatbot in Studio Pro, similar to Copilot, allowing developers to ask questions in real-time and receive suggestions tailored to their specific issues.

This feature enhances the development process by providing recommendations based on existing application models and industry best practices.

2.2. Assistance: Recommendations and Optimizations

In addition to its role as a guide, MAIA offers a set of recommendations tailored to the project's context. With its Best Practice Recommender, it can analyze code and propose improvements based on Mendix standards.

It also plays a key role in workflow optimization, suggesting structures best suited to the project's objectives. Additionally, it can identify the most efficient business logic and recommend optimized microflows, making it easier to set up a robust architecture.

Since user experience is also a crucial aspect of application development, MAIA provides UI adjustments and improvements based on UX best practices.

Finally, to ensure the quality and reliability of applications, MAIA enables the automatic generation of test scenarios to detect potential errors and improve the final product’s robustness.

2.3. Generation: Automation of Creation

Beyond guidance and recommendations, MAIA features automatic generation capabilities. It can propose translations for application texts using its Translations Generator, making multilingual application deployment more efficient.

Within the Mendix community, MAIA simplifies information retrieval by summarizing available questions and answers, allowing developers to quickly find solutions without manually browsing through discussions.

Finally, MAIA offers an advanced Domain Model Generator, which allows for the rapid structuring of a project based on business needs. This simplifies the design process and significantly reduces the time required to set up the application's database model.

3. Use Case: Creating a Ticket Management Model

3.1 Context

To illustrate MAIA's capabilities, we will use a generic example of a ticket management application. The goal is to allow users to create tickets, assign them to a responsible person, and track their progress through a status and priority system.

3.2 Accessing MAIA and Entering the Business Description

Business Description Provided to MAIA

A ticket must have a unique ID, a title, a description, a status (open, in progress, resolved, closed), a priority (low, medium, high, critical), and must be assigned to a user.

Steps to Follow

1. Open Mendix Studio Pro.
2. Verify that MAIA is activated:
   • Go to Edit > Preferences > New Feature.
   • Enable "domain model generator (experimental)".

3. Access the feature in the Mendix interface.

4. Enter the business description in the dedicated field for automatic model generation and start the generation process.

3.3 Defining the Generated Data Model

Result Obtained with MAIA

• Created Entities: Ticket, User.
• Automatically Assigned Attributes: TicketID (AutoNumber), Title (String), Description (String), Status (String), Priority (String) for Ticket.
• Established Relationships:
  • A Ticket is assigned to a User via the Ticket_AssignedUser relationship.
  • The Status and Priority values are managed as attributes within Ticket instead of separate entities.

3.4 Improving the Model with Adjusted Prompts

In the first iteration, MAIA generated a model where Status and Priority were simple attributes within the Ticket entity. However, to better structure the database, we adjusted our prompt to force MAIA to create these elements as separate entities.

Improved Prompt

A ticket must have a unique ID, a title, and a description.
A ticket must be assigned a status. The status must be stored in a separate entity called "TicketStatus," which contains predefined values: open, in progress, resolved, closed.
A ticket must also be assigned a priority. The priority must be stored in a separate entity called "TicketPriority," which contains predefined values: low, medium, high, critical.
Ensure that TicketStatus and TicketPriority are separate entities, not attributes.

With this adjusted prompt, MAIA generated a more structured model, including distinct entities TicketStatus and TicketPriority, along with the appropriate relationships.

3.5 Adding Missing Attributes to User

Although MAIA has now generated the TicketStatus and TicketPriority entities, the User entity still lacks attributes like Name and Email. We will therefore make one last prompt to include these elements.

Improved Prompt (Adding Attributes to User)

A user must have a unique ID, a name, and an email address.
Each user can be assigned to multiple tickets.

With this improvement, MAIA generated a model where User now contains the attributes Name (String) and EmailAddress (String).

3.6 Benefits and Limitations

Benefits

Using MAIA allows for the automatic generation of a structured data model in just seconds, significantly reducing the manual work required by developers. Furthermore, MAIA offers great flexibility by enabling modifications through prompt adjustments, ensuring a model that precisely meets the specific needs of the project.

Limitations

MAIA's efficiency heavily depends on the accuracy of the descriptions provided in the prompts. A vague or incomplete description may result in a model that does not fully meet business expectations.

4. Limitations and Future Perspectives

Although MAIA is a powerful tool, it has certain limitations. Its performance is heavily dependent on existing Mendix models, which can restrict flexibility when a project requires highly specific business logic.

Furthermore, while artificial intelligence helps automate many tasks, human validation remains essential to avoid errors and ensure that recommendations are suitable for the project's specific needs.

Since MAIA is an evolving tool, its capabilities are expected to continue improving over time with the integration of more advanced AI models and better customization of recommendations.

Next page → Marketplace