Agentic AI

The Agentic AI functionality is in its development stage and will only be available for a select group of beta testers. As soon as we feel comfortable, we'll make this feature available for general use.

This article serves as an overview of the integration and its associated tools. Subsequent sections will provide technical implementation details and deeper guidance.

Using AI and AI Agents can provide powerful features, but should also be treated very carefully - the AI "blackbox" is not an exact and predictable application. Use it wisely! Implement guardrails; keep monitoring, evaluating and improving the system.

Introduction to AI in the Modeler

Agentic AI in the WEM Modeler enhances your applications with intelligent automation through AI Agents. These agents can handle tasks such as processing user input, generating dynamic responses, and making data-driven decisions within workflows. By embedding AI into your no-code applications, you can create smarter, more responsive solutions with minimal complexity.

AI agents can analyze customer inquiries, summarize large datasets, create documents or execute business rules based on contextual understanding. They may improve user interactions, automate repetitive processes, and enhance decision-making, offering powerful capabilities that seamlessly integrate into your applications.

Understanding AI and Agentic AI

Agentic AI refers to AI systems that not only generate responses (like ChatGPT and the likes) but also take actions autonomously within a defined scope. These Agents can maintain conversation context, execute functions, and interact with external systems, allowing them to perform more complex tasks beyond simple input-output operations.

For a more in-depth understanding of Agentic AI and the differences from traditional AI models can be found on the OpenAI guide about Agents.

AI Capabilities in Your WEM Application

AI Agents in WEM Modeler can handle a wide range of tasks, from automating customer interactions and processing natural language inputs to generating data-driven insights and recommendations. Some examples include AI-powered chatbots, workflow automation, intelligent decision-making, and dynamic content generation. This is just a small sample — AI capabilities can be applied in many more ways to enhance applications.

Within WEM Modeler, AI Agents are seamlessly integrated into the no-code platform, orchestrating complex processes without requiring manual intervention. These agents can be configured and managed within the Modeler, enabling developers to create intelligent workflows that interact dynamically with users and data sources.

Risks & Best Practices

Potential Risks of AI in Applications

• Unpredictable behaviour: AI models may behave unexpectedly.

• Data privacy & security: Proper safeguards for handling sensitive data are required.

• Bias & fairness: AI models can inherit biases from training data.

• Performance impact: AI agents may introduce latency.

• Unpredictable costs: using AI is not free - you need a paid license and will pay for tokens used against the license.

• Environmental impact: AI features need heavy computational power - which needs electricity, and quite a substantial bit more compared to normal computer/datacenter usage without AI.

Best Practices for AI Integration

• Limit AI autonomy: Supervise AI actions to avoid unintended behaviour.

• Ensure transparency: Document AI usage clearly to better understand its actions and help with debugging.

• Optimize performance: Use AI selectively to prevent processing overhead and unexpected costs.

• Restrict waiting: Use Async task to ensure users do not have to wait for responses (WEM will make changes in future to make AI functions async by default).

• Data privacy & security: Proper safeguards and understand how your sensitive data is used.