The supply chain industry has never invested more in technology. AI, robotics, automation, visibility platforms: the roadmaps are ambitious and the budgets are real. Yet when you talk to operators honestly, the results are uneven. Some companies are pulling ahead. Others are stuck watching expensive deployments underperform.
So what separates the two? At MODEX 2026, we sat down with Shylaja Kadamby
AVP, Consulting Services, Everest Technologies and Rik Schrader, CRO, GreyOrange who have led hundreds of implementations, and learned the answer isn’t strategy or technology. It’s execution. Here are 10 hard-won lessons for making AI and automation actually deliver.
#1 Recognize that the execution gap is not new. It’s just more expensive now.
The gap between technology investment and operational results has always existed. What’s changed is the magnifying glass. As automation grows more complex and AI more prevalent, the cost of that gap grows with it: wasted investment, missed SLAs, failed deployments. Operators who acknowledge this upfront are far better positioned to close it than those who assume the next technology purchase will fix what the last one didn’t.
#2 Start with process clarity, not technology
Before a single robot ships or an AI model goes live, every stakeholder needs to understand how work flows through the warehouse, and why. Without process clarity, automation amplifies confusion instead of solving it. The technology will only ever be as smart as the process it’s built on.
“The underlying factor that unifies all of these things is process clarity. Without it, all of these advancements are going to fall flat.”
— Shylaja Kadamby, Everest Technologies
#3 Your WMS is not plumbing. Treat it like the engine it is, and fix it before you add AI on top.
A warehouse management system has never been passive. It makes hundreds of decisions every shift: labor allocation, task prioritization, inventory positioning, order fulfillment sequencing. AI doesn’t replace that decision-making; it depends on it. Operators who layer AI onto an aging, misconfigured, or neglected WMS quickly discover they’ve built something that confidently optimizes the wrong things. Upgrading the WMS and cleaning the underlying data is not a delay tactic. It is the prerequisite.
“If you want to add an extra bedroom to your house but the foundation is crumbling in the basement, I don’t know if that’s a smart move.”
— Rik Schrader, GreyOrange
#4 Make data quality a company-wide commitment, not an IT task
Dirty data is the most consistently cited root cause of underperforming deployments, and it rarely lives in one system. Data permeates an organization: WMS, ERP, automation controllers, order management. Getting to a single version of the truth requires alignment across every team that touches it, not just the implementation squad. Before go-live, data conversion testing deserves as much scrutiny as UAT. A mismatch in something as basic as unit-of-measure can bring throughput to a halt on day one.
#5 Think integration strategy from day one
There is no standalone technology in a modern warehouse. Every system, from WMS and WCS to AMRs, sorters, and conveyors, shares data and depends on other systems to function. Integration strategy consistently gets put on the back burner, and it consistently causes go-live failures. Map your data relationships early: who is the master, who is the child, and how does information flow in real time when exceptions occur.
#6 Simulate before you deploy, and make it a full picture, not a snapshot
Most deployment teams run some form of pre-go-live modeling, but too often it reflects a single, average operating day. Real warehouses don’t run on averages. Before committing to an automation or AI configuration, model what the environment looks like at peak, during a go-to-market shift, or when order and SKU profiles change significantly. The gap between what was simulated and what actually happens on the floor is one of the most consistent and avoidable reasons deployments miss their numbers.
#7 Manage scope like your project depends on it, because it does
Enterprise warehouse projects always have a want-versus-need tension. Declaring an ambitious vision is easy; executing it is disruptive and complex. The operators who succeed are the ones who commit to a disciplined phase-one scope, get it live, prove value, and then build from there. Trying to build the perfect warehouse in one deployment is one of the most reliable ways to deliver a disappointing one.
#8 Bring floor-level users into the design process early
The most common change management mistake is not involving the right people at the right time. Frontline workers and supervisors should not encounter a new WMS for the first time at UAT. They need to be in design sessions: seeing what is changing and why, and shaping the outcome they will live with every shift. Workers who help build the system are far more likely to own it at go-live, and far less likely to work around it.
“Gen Y and Gen Z workers in warehouses today don’t like to be just told what’s changing. They like to be involved. Get them into the journey.”
— Shylaja Kadamby, Everest Technologies
#9 Rethink what ‘AI delivering value’ actually looks like
The most overhyped AI application in warehousing today is the fully autonomous, lights-out operation. That is years away for most environments. The most underrated is invisible, real-time optimization: labor reallocation when a worker calls in sick, task reprioritization when an order changes mid-belt, predictive congestion routing that prevents bottlenecks before they form. The best AI outcome is the problem nobody had to solve manually. That is where the real ROI lives right now.
#10 Bring AI in Smartly, Not Evangelically
There are two ways operators are getting AI wrong today: those too cautious to start, and those implementing it everywhere without understanding what it actually requires. Neither approach wins. The operations that pull ahead will be those that identify specific, high-frequency decisions where AI can drive measurable improvement, then build incrementally from a solid execution foundation rather than betting everything on a single sweeping transformation.
The Bottom Line
The warehouses winning right now are not the ones with the most advanced technology. They are the ones with clean data, a well-configured WMS, disciplined execution, and a workforce that owns the systems they use. Technology is only a multiplier. Applied to a weak foundation, it multiplies the problems just as readily as the gains.
Fix the foundation. Close the execution gap. Then let AI do its job.
