AI-Assisted Development in Action with Mendix

How is AI-assisted development different from traditional software development?

AI-assisted development is the use of machine learning to assist developers by providing recommendations in real-time and enforcing best practices throughout the entire software development lifecycle (SDLC).

AI has the ability to provide guidance, assistance, and generation throughout the entire software development lifecycle (SDLC).

AI gives guidance by enabling developers to ask the IDE development-related questions. In turn, the IDE gives immediate, insightful responses with access to the sources used to derive the answers.

AI can assist with making real-time recommendations and enforcing best practices while creating logic, or automate remedial tasks that often slow down developers using traditional development methods.

A major difference between AI-assisted development and traditional is using generative AI. Generative AI can accelerate the development process and enhance consistency and quality by:

• Producing front-end and back-end logic
• Creating application components
• Automating tasks like creating SQL commands, and test data, for example

What features does Mendix offer for AI-assisted development?

The work we’ve done with AI gives you and your development teams more guidance around software development.

With that said we’d like you to meet Maia, or Mendix AI assistance. Maia is our suite of AI-assisted development tools built into the Mendix Studio Pro IDE.

The services you find under Maia allow you to get the speed you need to deliver solutions and start generating value more quickly. They’ll also give you a way to gain higher quality right out of the gate. The tools you find in Maia are all about intelligent guidance, assistance, and generation

Maia Chat

Our previous chatbot, Mendix Chat, has now been integrated into Maia and is now known as Maia Chat. Maia Chat provides immediate, detailed answers on Mendix-related questions and is powered by Generative AI.

We see Maia Chat benefitting new users of the Mendix platform to get up to speed more quickly and helping those with Mendix experience get answers more quickly.

More updates to Maia chat include a 40% faster response time, copy/paste shortcuts, and enhanced readability.

The improvements we’ve made to Maia Chat are vital to helping your developers build faster, which leads to getting solutions into production more quickly, and creating value as fast as you need.

AI-assisted development: Maia Recommenders

Also found under the Maia umbrella are a number of recommenders. Recommenders provide real-time and context-driven next steps or best practices so that you can complete your build faster.

If you’ve used Mendix 10 at all, you’ve seen this before with our Logic and Best Practices Bots. These are now known as Maia Logic and Maia Best Practice Recommenders.

The Logic Recommender provides real-time and context-driven options for the next most relevant actions when building microflows and nanoflows. New with 10.12 is the Workflow Recommender. Workflow Recommenders is like the Logic Recommender in terms of UI, but suggests user tasks and workflow elements like timer events, decisions, splits, to name a few) which suggests logic for business processes.

Generative AI for development

Our foray into the world of generative AI starts with two generators found in the IDE. The first is the Domain Model Generator. This won’t be ready until 10.13, but we’re just so excited, we can’t wait to talk about it.

The Domain Model Generator enables you to converse with Maia in the IDE to build a domain model. Using prompts like, “Maia, I want to build a bike rental application” and “there will be bikes and customers,” Maia will create entities and establish the relationship between them to generate a model.

With 10.12, we’re launching the Maia Translations Generator, translating text for multi-lingual applications. With one click from the batch translate window, the Translations Generator efficiently translates all untranslated texts in your app across a wealth of different languages and ensures no text is left untranslated.

In the months ahead, we’ll have more ways to assist developers. But that doesn’t mean we’re done talking about generative AI.

Build AI-augmented applications with low-code

The benefits of AI aren’t just for developers. We want you to build smart software that moves the needle in terms of business value. This is why we continue to build out tools and make partnerships so you can deliver AI-augmented applications more easily than with traditional development.

Mendix makes it easy for you to use large language models and natural language processing within your applications. Here are a few ways we’re doing that in 10.12.

Conversational UI

Customer experience is everything. This is why building chatbots and other conversational-type experiences should be top of mind for most enterprises.

With Conversational UI, you get reusable components that let you more easily build these types of experiences, whether it’s a standalone chatbot or a conversational experience integrated into your Mendix application.

Depending on your preference, you can use the models from either the OpenAI or Amazon Bedrock connectors and leverage your own data through retrieval-augmented generation (RAG). Both connectors have been updated to be compatible with Conversational UI.

AI Starter Apps

Want to build an enterprise-grade ChatGPT-like solution, but need it to be private? We’ve got you covered with our AI Bot Starter Application. With this, you can build a bot in minutes that will give your users

• Someone to brainstorm with
• A copywriter to put together emails and social posts
• An assistant who can summarize large amounts of text
• A researcher who can analyze PDF documents
• A developer who can help with logic challenges

This bot is compatible with Azure OpenAI and Amazon Bedrock and supports linking it to your data sources in a RAG setup.

We’re also providing a Blank GenAI Starter App which accelerates the creation of your own custom GenAI project.

AI and low-code belong together

When it comes to any new technology, we want to see how we can use it to help you, our valued customers. We also want to see how we can help you leverage it, too.

We will continue developing and refining our offerings around AI-assisted development and AI-augmented applications to make sure that you can deliver value as fast as possible.