Lucas Systems and Carnegie Mellon develop warehouse prediction models

By Amara Osei

A spring 2026 Carnegie Mellon University master’s cohort supplied Lucas Systems with prediction models developed after studying hundreds of data attributes tied to warehouse users and behavior. The project forms part of a newly announced research partnership with the warehousing and distribution-center software technology provider.

The students belong to the spring 2026 cohort of Carnegie Mellon University’s Master of Science in Applied Data Science. Lucas Systems says the partnership could have commercial impact by finding new ways for supervisors to use data analysis for more accurate, productive workforce decisions.

More data, fewer usable answers

Lucas Systems says operators collect abundant warehouse data but still lack the tools to use its insights efficiently when managing labor and tasks. The company also argues that operational complexity and constant observation make exclusive reliance on human decision-making impractical.

That is a broad diagnosis: a dashboard can display plenty while declining to answer the next staffing question. Better analysis, Lucas Systems says, could help supervisors allocate work with fewer decisions resting on intuition.

Kent Szalla, Lucas Systems’ vice president of software engineering and customer success, said: “Warehouse managers are drowning in data but starving for insight. They’re making million-dollar decisions based on experience and gut feel because the data they have isn’t organized in a way that tells them what to do next.”

He continued: “What excites me about this research is that it points toward a future where the system surfaces the right signal for planning at the right moment. Supervisors aren’t guessing, they’re deciding.” That puts the intended commercial benefit in operational terms: timely guidance, not another retrospective chart.

From course work to prediction models

The students analyzed hundreds of attributes related to warehouse users and behavior before providing their prediction models. Lucas Systems’ data scientists and software engineers say those perspectives could move planning beyond guesswork and experience-based habits.

André Lessa, Lucas Systems’ director of R&D, said: “They collaborated with us on the very complicated math. As a result, we have a strong foundation to realize our vision.” The statement credits the students’ modeling work without specifying its current accuracy.

The joint project is one of several ongoing collaborations between Lucas Systems and Carnegie Mellon University. The announcement does not say whether this model set has entered a commercial product or a customer pilot; for buyers, this remains promising research rather than a line item they can price.

What changes for operators?

For warehouse operators and shippers dependent on fulfillment centers, the potential benefit is clearer labor and task allocation, with productivity gains and fewer decisions made from gut feel. The announcement offers no measured productivity improvement or labor-hour saving, so the gain remains a possibility rather than a demonstrated outcome.

For carriers and freight forwarders, the announcement identifies no direct transport change. It names no trade lane, TEU or tonnage movement, capacity percentage, transit time or freight rate, which means nobody should attach a transport-market benefit to this research yet.

Nor does Lucas Systems disclose a named deployment site, model target, implementation timetable or commercial terms. Its technical team describes the work as potentially transformative, but the announcement offers no metric to test that claim; the label may travel faster than the evidence.

The announcement also carries a diary note: IntraLogisteX will return to the United States on September 15-16, 2027, for supply chain, logistics, warehousing, fulfillment, manufacturing and automation professionals. The event listing sits outside the research announcement and offers no new information about the models.

The trajectory is toward planning software that identifies the right signal at the right moment and reduces reliance on guesswork. For carriers and forwarders, no capacity or freight-rate trajectory is established; for operators, the next test is whether the models can turn abundant data into better workforce decisions.