A Dematic‑hosted Innovation Roundtable explores where AI can support sustainability and circularity goals — and why human judgment, oversight, and accountability remain essential.

AI can support better sustainability and circularity decisions – but only when organizations are clear about what they are trying to solve. That message anchored a Dematic‑hosted Innovation Roundtable focused on Innovation & AI for Sustainability & Circularity, featuring Geertrui Mieke De Ketelaere.

A recognized expert in sustainable, ethical, and trustworthy AI, De Ketelaere, an adjunct professor at Vlerick Business School, works closely with organizations eager to adopt AI — but often before defining clear objectives. Whether driven by hype, competitive pressure, or fear of falling behind, many companies move too quickly to implementation without first understanding the sustainability or circularity challenge they want to address.

“AI can create value, but it has to start with objectives,” De Ketelaere explains “How far you go depends on the industry, the process, and the stage of maturity you’re in.”

That distinction matters. Sustainability efforts typically focus on reducing environmental impact across operations, while circularity aims to extend the life of products and materials to minimize waste. In both cases, AI can play a role — but only when applied deliberately, with a clear understanding of what success actually looks like.

The discussion also highlighted a broader challenge organizations face today: separating meaningful AI applications from inflated expectations. De Ketelaere emphasized that governance, operational readiness, and human oversight are critical in making that distinction.

Without those foundations, even well‑intentioned AI initiatives can lead to questionable sustainability claims. Measuring progress too narrowly — such as pointing to waste reduction or throughput improvements alone — does not automatically indicate real or lasting impact if broader operational, environmental, and business effects are ignored.

“Don’t be blinded by numbers or statistics alone,” De Ketelaere cautions. Organizations need to understand their own level of maturity and context before drawing conclusions about success.

That perspective aligns closely with how Dematic approaches automation and intelligence more broadly: performance and sustainability are not defined at a single moment in time, but over the full lifecycle of an operation.

Many organizations want AI that supports their sustainability goals, but some still start in the wrong place. They begin with the technology, move quickly toward implementation, and only later ask what operational, environmental, or circularity problem they’re trying to solve.

“They rush it a little bit because there’s a lot of hype,” De Ketelaere says. “You really have to first think, what do you want to reach?”

That question matters because AI doesn’t fix weak objectives, poor data, or immature processes. It can help organizations find patterns and support better decisions, but only when the foundation is ready. If a company wants to train an AI system on its own operational data, that data must be accurate, complete, and useful.

A lower waste rate, higher throughput number, or better-quality score can make an AI initiative look successful. But when companies use AI to support sustainability goals, they need to understand what those numbers do and don’t prove.

De Ketelaere also warned companies against measuring sustainability gains too narrowly. For example, waste reduction or output improvement doesn’t automatically make an AI initiative successful if the organization ignores the broader operational and business impact.

“Don’t be blind to the numbers and statistics,” De Ketelaere cautions. “Companies first need to understand their own level of maturity.”

That’s why companies need to look beyond the strongest number in the report. A system may, for example, reduce waste, improve quality, or help teams plan maintenance more effectively, but those gains still must be weighed against the full operational impact. That includes the effects on sensors, data infrastructure, energy use, employee training, process changes, and the people running the operation.

Circularity pushes companies to keep products, parts, and materials in use longer, which means thinking earlier about design, materials, and recovery. It asks how materials get selected, how products get made, how long they remain useful, and how they can reenter the economy instead of becoming waste.

De Ketelaere pointed to a common reason companies miss opportunities: they keep using the same materials, designs, and processes. “We sometimes do things because we’ve always done them this way,” De Ketelaere says.

AI can step in and support early design decisions by helping engineers consider alternative materials, components, or product approaches that may support a longer lifecycle or easier recovery at end-of-life. Engineers still decide what works, but AI can bring more options to the table.

This article reflects Dematic’s role as a connector to outside expertise on issues shaping business, operations, and automation. It also reinforces a simple but powerful reality: The Mind Behind the Machine has to come before the machine itself.

Technology continues to evolve quickly. Understanding — grounded in experience, judgment, and accountability — is what turns that technology into lasting performance and meaningful progress.