On May 8th, Somia Solutions had the pleasure of participating in the professional session “Transforming Risk Prevention and Occupational Health through AI”. The event brought together professionals from the occupational health and safety sector to explore the opportunities that Artificial Intelligence can bring to prevention.

The event, made possible thanks to the warm invitation from Salut i Treball and its CEO Paco Lari Carrillo, allowed us to connect directly with professionals working in health and safety. We were also honored to count on the support of the Cambra de Tarragona, especially President Laura Roigé, Vice President Lluís Colet Roselló, and Communications Director Jordi Salvadó Folch, who made everything run smoothly.

From Theory to Practice: AI with Common Sense

We started by laying the groundwork with a clear and engaging explanation of what AI really is—cutting through the hype to focus on what matters to companies. We explained the difference between traditional AI and Generative AI, and then took it a step further to explore how Large Language Models (LLMs) like ChatGPT operate.

Understanding AI Agents

We also introduced a crucial distinction:

• A Language Model is powerful, but reactive.

• A Tool-augmented Model can retrieve data or perform external tasks but still it is at the developing stage, not ready for the real-world cases.

• But an AI Agent goes further—following its own flow of reasoning and interacting continuously with its environment, adapting to user needs.

To make this distinction more tangible, we used the allegory of a student:

• A Language Model is like a student who has read thousands of books and can answer questions based on memory. It’s reactive and depends entirely on the input.

• 

  A Tool-augmented Model is like giving that student access to a teacher who explains how to use tools (like calculators, encyclopedias, or search engines). The student can now retrieve and apply information—but still requires guidance.

• 

  An AI Agent, in contrast, is like giving the student a planner, access to tools, and the autonomy to make decisions. This student doesn’t just respond—they plan, adapt, correct errors, and follow a process independently to reach a goal.

This is precisely what Somia enables: giving companies—especially SMEs—the power to design and control their own AI agents without the need for programming. These agents can follow structured workflows, interact with tools, and adjust their behavior based on the needs of the business.

At Somia, our mission is to democratize AI by enabling any SME to create and operate their own AI agents—without needing to code, and with full control over how the agent behaves and learns.

We illustrated these ideas with real-world examples showing how agents are being used today in prevention services: from automating documentation to improving internal knowledge access and client interactions.

AI with Control, Validation, and Responsibility

Here we tackled the big questions:

We introduced our approach to AI Readiness, which evaluates the maturity of a company before implementing AI, ensuring the solution fits their context. We also offered practical recommendations for starting a pilot AI project with minimal risk and maximum impact.

Most importantly, we showed that building trustworthy AI isn’t just for tech giants. With the right support and the right tools—like the Somia platform—

If you couldn’t attend, we’ve got you covered:

✨ Want to see it in action? Contact us for a live demo and let’s explore how AI can support your team today.