Warehouses are turning to digital twins because traditional planning tools cannot keep up with the level of volatility they now face. Recent years brought extreme swings in demand, supplier shifts, labor shortages, and transportation disruptions.
Digital Twins Game and Gaining Advantage in Warehousing from using AI
Q&A with Matt Derganc, Senior Director | SSA & Co.
Tell us about yourself and your expertise.
I help companies strengthen supply chain performance, manufacturing operations, and enterprise systems to deliver measurable, durable gains. Over more than twenty years, I have converted strategy into results across the middle market, splitting my time evenly between operating roles with direct accountability and consulting engagements with public, private, and private equity owned businesses.
My work spans more than forty organizations in heavy industrials, consumer goods, technology, energy, logistics, and retail, and includes system modernization efforts, integration of manufacturing operations, long-term manufacturing strategy design, and programs that improve working capital and operating performance.
How and why are warehouses starting to use digital twins?
Warehouses are turning to digital twins because traditional planning tools cannot keep up with the level of volatility they now face. Recent years brought extreme swings in demand, supplier shifts, labor shortages, and transportation disruptions. Many teams were left making multi-million-dollar decisions with spreadsheets that were not designed for rapid, high-stakes scenario testing.
Digital twins allow operators to convert this volatility into a planning advantage. They simulate shift schedules, wave plans, and throughput decisions before execution. Companies with twins turn volatility into advantage; companies without them are left reacting.
Where does AI provide real advantage today?
AI delivers the most value in the daily operational decisions that drive labor productivity. Up to a third of warehouse labor is lost to inefficient staffing, long pick paths, and inconsistent slotting decisions. A twin enhanced with AI ingests the live order mix each morning, recommends how to rebalance labor by zone, and identifies travel-time reductions faster than any manual analysis.
Planning teams use AI to test wave sizes, batching rules, and routing choices throughout the day. The impact shows up almost immediately in lines per hour and travel distance per order. This is where AI consistently earns a measurable return.
What efficiencies can they find, and can this lead to organic growth?
Digital twins typically unlock 10 to 25 percent labor productivity improvement when used in daily planning. Overtime costs, temporary labor, and error rates often decline as a result. Space utilization improves as slotting becomes more systematic, and throughput increases without adding square footage.
These gains accumulate. Instead of seeing improvements fade each quarter as processes drift, twin-driven operations continue to refine performance. Facilities effectively create new capacity inside their existing footprint, which allows them to grow without immediately investing in new buildings. This is organic growth fueled by operational discipline rather than capital expenditure.
How is capacity optimized using digital twins?
Digital twins allow teams to pressure-test capacity decisions before they face real constraints. Retailers model peak season surges months in advance and identify the specific zones that will bottleneck. E-commerce operations test how same-day delivery commitments will influence flow and resource needs.
At the network level, twins evaluate proposed facilities and often uncover issues that are not visible in static spreadsheets. Avoiding even one underperforming site represents a significant capital savings. The rule is simple: simulate major decisions before executing them.
What are the biggest hurdles to deployment, especially the human ones that technology alone can't solve?
Weak processes kill more twins than bad tech ever will. Many implementations struggle because foundational processes like scanning, exception handling, and standard work are not performed consistently across shifts. A twin relies on accurate, complete, and timely data, so weak process discipline undermines the model immediately.
Culturally, supervisors and operators may hesitate to accept recommendations from a system they did not help design. Many have seen past technology projects create more work, not less. Trust is earned through transparency and shared ownership. Teams need to see how recommendations are produced and how their operational realities are reflected in the model.
How does the workforce realign itself to a digital twin environment? What skills matter more in that environment?
The workforce shifts from reactive problem solving to consistent execution and proactive pattern recognition. Operators who were valued for improvising during busy periods now become most effective when they follow optimized processes that allow the model to perform accurately.
Supervisors transition from directing traffic to monitoring exceptions and validating why the model recommended specific actions. Data literacy becomes an essential skill. The most valuable capability is translating between model logic and real-world constraints, then helping refine the system as operations evolve.
What lets companies move fast with this kind of transformation without breaking operations?
Speed comes from focus. Successful companies begin with one clean and well-understood pilot. They run the digital twin in parallel with real operations for several weeks, using this time to compare decisions and verify accuracy before making any operational changes. This approach demonstrates value, builds credibility with frontline teams, and exposes data or process gaps early.
After proving impact on one flow, the organization expands to adjacent processes. This controlled progression prevents operational disruption and accelerates adoption.
What role does data discipline play in separating leaders from laggards?
Data discipline is the defining difference between companies that succeed with digital twins and those that do not. A twin cannot function when location accuracy is inconsistent, scan compliance is incomplete, or inventory accuracy falls below required thresholds. Even small data gaps cascade into incorrect recommendations.
Common integration challenges include timestamp inconsistencies, duplicate transaction IDs, missing movement scans, phantom inventory from canceled orders, and delays in order-status updates. Leading organizations address these issues before launch and establish clear ownership for data standards and governance. Laggards discover them only after the model begins producing conflicting results.
What does "trusting the model" look like in practice for supervisors, planners, and operators?
Trust develops when the logic behind recommendations is visible and consistently validated. Supervisors need to understand why the model suggests reallocating labor at certain times. When the reasoning aligns with real-world conditions, they begin to rely on the guidance.
Planners gain trust when predicted outcomes match actual performance and when exceptions are surfaced quickly. Operators develop trust when following the recommended pick path or workflow makes their jobs easier. The goal isn’t blind compliance; it’s a partnership in which humans understand when to follow the model and when to override.
How do you involve frontline teams early so the technology becomes useful instead of threatening?
Make them co-authors, not end-users. An effective approach is to designate shift leads as "twin champions" who explore simulations, validate assumptions, and identify constraints that the model must respect. When real operational knowledge is incorporated into the twin, employees see themselves reflected in the system rather than replaced by it.
Early involvement removes the mystery behind the technology and builds a sense of ownership. Teams are far more likely to embrace tools they helped shape.
What early behaviors signal that a workforce is adapting well? What signals suggest the opposite?
Positive indicators include supervisors referencing simulation results during shift meetings, operators asking what the twin recommends before escalating issues, and frontline suggestions flowing regularly into model refinement. These behaviors signal that the workforce sees the twin as a helpful decision tool.
Warning signs include declining scan compliance, informal workarounds that bypass the system, and supervisors maintaining separate schedules outside the model. The most concerning indicator is when operators stop reporting exceptions because they believe the system cannot process them. At that point, the twin becomes fiction.
The content & opinions in this article are the author’s and do not necessarily represent the views of ManufacturingTomorrow
Featured Product
