Food and Beverage businesses will transform their operations with agentic AI* – but won’t realize its full potential until AI earns humanity’s trust, according to Ben Dorks, CEO of Ideagen. Dorks said the sector’s compliance obligation is the one domain where autonomous AI will thrive – despite leading analysts warning 40% will fail in the next two years.
Speaking at the launch of their own agentic AI platform Ideagen Mazlan, Dorks, said: “Autonomous AI will succeed in compliance because failure is measurable, stakes are high and outcomes are binary.
“AI that continuously monitors jurisdictional or industry specific compliance frameworks, detects patterns across thousands of records, raises actions for others to follow, flags gaps before audits, autonomously? That’s agentic. It will literally save lives and is the most transformative shift in protecting people, products and processes we have ever seen.
“We could have gone further, but the world isn’t ready for that yet, agentic AI still needs to earn humanity’s trust before we give it full control. But what it can do is to keep workplaces safe, reduce product recalls, keep supply chains moving, prevent harmful contaminants entering food”
Initial piloting has demonstrated significant acceleration in adoption and benefit, in some cases squeezing 30 minute-tasks down to just two minutes. Other estimates show that enterprise level implementation that could take a company six to nine months, could be achieved in around 30 days.
“In the new year, we’ll be adding things like a voice interface so workers on the frontline, who might be in heavy PPE, can just tell Mazlan what’s happened,” said Dorks.
“And it’s intuitive, people won’t need to be trained to use it. If we can break down barriers to the adoption of digital solutions, then businesses are operating with better data which in turn results in the safer products and processes and a better return on that investment.”
Dorks was also quick to stress that Ideagen Mazlan is not a bolt-on solution to increase productivity: “It’s intelligence, built in. Most ‘agentic AI’ isn’t agentic at all – it’s assistive AI with better marketing! AI that accurately answers questions when you ask them? That’s a chatbot.
“Our agentic AI works on three simple principles: it’s built into Ideagen solutions, it understands your specific industry context and appears naturally throughout your workflows.
“It knows OSHA, ISO and the legislation in your jurisdiction, it knows your own policies and data then uses this to analyze incidents, identifying risks, automating compliance tasks.
“But crucially, like autopilot technology on planes, we’ve built it to ensure that the power still sits with the human. Final oversight and approval remains with the quality managers and health and safety teams, but Mazlan is autonomously doing the heavy lifting.”
* Editors Note: Agentic AI refers to advanced AI systems that operate autonomously, setting goals, planning, and executing complex, multi-step tasks with minimal human intervention, going beyond simple command-response or content generation to proactively solve problems and adapt to dynamic environments by using tools and collaborating with other agents. These systems are goal-oriented, adaptable, and capable of independent decision-making, enabling them to handle intricate workflows in fields like finance, healthcare, and logistics, turning AI from a tool into an initiative-taking partner.



