As digital tools proliferate at unprecedented speed—from advanced planning platforms and AI/ML engines to automation suites and digital twins—leaders face a fundamental question: Which technologies will genuinely move the needle? The modern supply chain is no longer about adopting the latest system but about orchestrating a cohesive tech stack that enhances visibility, resilience, and financial performance. This Special Report examines the critical first steps organizations should take before committing to new systems, the blind spots that derail early decision-making, and the deeper ROI levers that often go unmeasured. It also unpacks the evolving realities of AI/ML adoption and the shifting relationship between digitization and automation—setting the stage for expert perspectives on What Truly Delivers Value Today.
Over the past few years, digitization and automation have dominated boardroom discussions. But how exactly have they evolved? Have they progressed hand in hand, or have they developed in isolation? Can we realistically view them as parallel streams, or must they be treated differently?
Rajat Sharma
Rajat Sharma, Vice President – SCM & CX, Hamilton Housewares (Milton): I’d build on that by saying digitization is more than just an enabler—it’s a prerequisite. Unless you first translate a process into a digital model, automation cannot be meaningfully applied. Otherwise, what you’re left with is not automation but mechanization—machines that amplify human effort but remain dependent on it. Think of the difference between a worker operating a molding machine and a fully integrated CNC system. In the former, efficiency is scaled up, but the process remains human-driven. In the latter, the entire operation is digitally governed. That’s the leap digitization makes possible. So whether we’re talking about warehouses, shop floors, or production units, the sequence matters. Digitize first, then automate. Without that digital foundation, automation cannot deliver the intelligence, consistency, or scalability that businesses today demand.
Digitization and automation are not isolated journeys but interdependent stages of transformation. Digitization lays the groundwork—it converts processes, assets, and flows into structured data. Automation then builds on that foundation, adding intelligence, efficiency, and scalability. The degree to which organizations pursue each depends on their maturity, their business priorities, and above all, the ROI they seek. The sequencing may vary, the adoption curve may be uneven, but the direction is clear: digitization and automation are not parallel roads—they are two steps of the same staircase. One establishes the structure; the other helps you climb higher.
Shreyas Dhore
Shreyas Dhore, GM – Supply Management Shared Services, John Deere: I would nuance the discussion by pointing out that while digitization and automation often appear to move in parallel, in reality, there is always a lag. Automation cannot function in a vacuum—it is entirely dependent on digitized data. Unless information is first captured, structured, and made accessible in digital form, automation cannot deliver its promise.
On the shop floor, this dependency is not always visible. When we see AMRs or ASRS systems operating, they look like they are running in perfect synchrony with the larger system. Yet behind that smooth operation lies a sequencing logic: the information has already been digitized before the automation layer can act on it.
In more complex environments such as supply chain process automation, the lag becomes far more evident. Robotic Process Automation in its earlier avatar, and even the intelligent process automation we see today, all share the same principle: automation is only as good as the digital foundation beneath it. If the data pipeline is fragmented, automation slows, stalls, or produces limited value. So while the technologies may feel concurrent, in practice digitization always sets the stage. Automation follows, adding efficiency and intelligence only once the groundwork of digitization is firmly in place.
Tannistha Ganguly
Tannistha Ganguly, Associate Director – Supply Chain WMS (IT), Kimberly-Clark: I see automation as a subset of digitization. Digitization is the overarching framework that makes everything else possible—it enables automation, artificial intelligence, RPA, and advanced analytics. The two aren’t competing priorities; they move together, but at a pace dictated by an organization’s maturity. In our warehouses, for instance, we run ASRS systems that feed data into our ERP and WMS. Alongside this, AMRs and AGVs operate autonomously while exchanging real-time, bidirectional data with our central systems. None of this exists in isolation—digitization provides the backbone, while automation extends its value.
But the real question is: how much of each do you actually need? Sometimes a simple ASRS solution delivers enough efficiency without requiring fleets of AGVs. Similarly, not every business benefits from advanced AI-powered analytics; in many cases, basic reporting is sufficient. It always comes back to ROI. Your level of digitization and automation should align with business needs and the value you intend to unlock, rather than chasing technology for its own sake.
When it comes to technology stacks, the choices can feel overwhelming. How should organizations think about this landscape and decide what’s right for them?
Rajat Sharma: The starting point is to recognize that no technology is inherently good or bad—each has its place. What really matters is 'HOW you choose' and the journey that decision takes. The tech stack today spans a wide spectrum: on one end you have Excel, and on the other, complex ERPs, SaaS platforms, mechanization, and even robotics. To make sense of this expanse, we classify it into four layers. The first is transaction systems—your operational backbone where activities are created, tracked, and recorded, such as ERPs or aggregation tools. On top of that sits the analytics layer, which transforms raw transaction data into insights you can actually use. Then come the decision systems, which deliver the most business value by moving from insights to action through optimizers, aggregators, or digital twins. Finally, there are action triggers, where technology meets the real world through IoT, robotics, and man–machine interactions. Now, the critical question is not just which layer to invest in, but how to sequence your choices. A common mistake is rushing to advanced decision systems without first stabilizing your transaction and analytics layers. Another is chasing “shiny” technologies that don’t align with your business needs. The smarter approach is to map your problems, place them against this four-layer framework, and then decide where the greatest value lies. In other words, don’t just ask, “What’s the best tech?”—ask instead, “What’s the right tech for where we are in our journey?”
When organizations look at technology, should they focus only on solving the immediate problem with the right tool—or is there a bigger picture to consider before investing?
Neha Singh
Neha Singh, VP Global E2E Supply Planning Transformation, Diageo: That’s the real crux of the digital transformation challenge. Too often, the instinct is to say: Here’s a problem, let’s find the best tool to fix it. I’ve seen this across industries and roles in operations, manufacturing, procurement, and planning. The pattern is the same: firefighting with technology. But that mindset rarely creates sustainable impact.
The smarter approach is to start with your digital architecture, not the tool. And you don’t need an expensive consultant to begin; your own team can do it if you bring the right people together. The first step is to map your taxonomy—your processes from Level 1 down to Level 5 or 6. Once that’s clear, create a capability map against it: where do we just need to be functional, where do we aim to improve, and where should we benchmark against the best in the world?
Now, once you have that map, ask the harder question: Do we really need a digital intervention here, or will process redesign suffice? Only if the answer is digital do you start looking at tools. And when you do, the choice becomes strategic: if service levels in planning are the issue, for instance, then you evaluate platforms like OMP, O9, Kinaxis, Aera, Blue Yonder etc.—but always in the context of your taxonomy and long-term digital ecosystem.
This sequencing is critical. By doing the groundwork in-house—taxonomy, capabilities, and priorities—you make sure you’re not just solving for today’s fire. You’re future-proofing. You’re asking: will this tool work with my ERP, support my digital agenda, and integrate with what’s coming next? That’s how you avoid fragmented systems and build a coherent, scalable digital foundation.
So my advice is simple: don’t start with the tool, start with your architecture. Do the taxonomy and capability mapping first, and then go to market with clarity. That’s when technology becomes a strategic accelerator, not just a patch.
Today, supply chains operate across multiple platforms—ERPs, specialized software, and in-house tools—and most of them aren’t integrated. Before investing in new technology, should companies focus on bringing it all together? How should they approach this challenge?
Tannistha Ganguly: Absolutely. This question reflects a broader shift in mindset. It’s no longer “Do we need technology?” The question now is “How do we manage it effectively and make it truly valuable?” Companies have adopted various digital solutions over the years— from large ERP systems forming the backbone of operations, to smaller, niche software addressing specific functional needs, to highly customized in-house applications built for critical business functions. Each system has value, but if they aren’t integrated, they become silos. And silos, no matter how sophisticated, create inefficiencies, duplicate efforts, and reduce the agility of the supply chain.
The first step, from an IT perspective, is to design an integration architecture before implementing new tools. Think of it as building the digital backbone of your organization. This backbone ensures that every system— from ERP to specialized third-party software to in-house applications—can communicate seamlessly. You choose the integration methodology that suits your organization—whether MuleSoft, SAP CPI, Node.js, or another technology— but the principle is the same: create a scalable, robust, and flexible framework before layering on more tools.
Why is this so critical? Let’s take an example: managing third-party logistics providers (3PLs). A company may work with dozens—or even hundreds—of 3PLs across geographies. Without a strong integration layer, you’d need to build hundreds of point-to-point connections to exchange operational data. Not only is this expensive and resource-intensive, it’s also extremely hard to maintain. With a well-designed integration backbone, you integrate once and scale infinitely. The same principle applies to warehouse systems, vendor platforms, and specialized analytics tools—once the backbone is in place, additional tools can plug in smoothly without creating chaos.
Integration also sets the stage for advanced digital initiatives, from predictive analytics to AI-driven decision-making. When systems are connected, data flows consistently, governance is easier, and insights can be trusted. Without integration, even the most advanced technologies fail to deliver impact because they are feeding off fragmented or inconsistent information.
In short, prioritize building your integration stack before investing in additional technology. It’s the foundation that enables flexibility, reduces operational complexity, and ensures that every technology investment is both strategic and scalable. Once this backbone is established, adding new capabilities— from analytics to automation—becomes far simpler, more reliable, and ultimately, more valuable for the business.
When organizations start exploring technology for their supply chains, what should be the initial steps? And where do they often go wrong?
Shreyas Dhore: That’s a crucial question because it goes to the heart of how technology adoption in supply chains succeeds—or fails. Let me begin with a quick backdrop. I’ve spent over two decades in supply chain leadership roles, including 14 years at John Deere and earlier stints with General Motors and Siemens. Over this period, I’ve witnessed the tech stack evolve from basic transactional systems and spreadsheets to ERP platforms, advanced analytics, IoT, robotics, and now AI-driven decision systems. The landscape has changed dramatically, but one common mistake persists: organizations often start with the technology rather than the problem.
Too often, teams get excited about a new tool and rush to implement it, hoping it will be a silver bullet. But what I’ve consistently seen is that when you target technology first, you risk misalignment—you may end up with a solution in search of a problem. The far more sustainable approach is to begin with a problem-first mindset. Ask: What is the pain point? What challenge are we trying to solve? How does this link to the broader supply chain strategy? Only then does the choice of technology start to make sense.
To make this more practical, I rely on what I call the Four A’s Framework for Technology in Supply Chains. It’s a simple but robust way to evaluate whether a solution truly adds value.
Applicability – This is the starting gate. Is the problem worth solving, and is technology the right lever to solve it? Not every challenge requires a digital solution; some may need process re-engineering or capability-building instead. Asking “Is it applicable?” keeps us grounded.
Affordability – Once you know the problem is real and relevant, the next question is about value. What benefits are we trying to unlock, and does the return justify the cost? Affordability isn’t only about budget—it’s about ensuring the solution delivers measurable business outcomes.
Accessibility – This is often overlooked but absolutely critical. Even the best-designed systems can fail if they aren’t usable by the people on the ground—planners, buyers, warehouse operators, inventory analysts. Technology has to be intuitive, easy to adopt, and integrated into daily workflows. Otherwise, it stays in pilot mode or becomes shelfware.
Availability – Finally, a solution must scale and sustain itself. Can it be reliably deployed across the organization? Will it continue to deliver value once the initial excitement fades? True availability means the technology is embedded into the fabric of operations and becomes self-sustaining over time.
When organizations ignore these steps, they often stumble. I’ve seen proofs of concept that looked brilliant on paper but collapsed in execution because they weren’t accessible or scalable. I’ve also seen expensive tools underutilized because the problem they were meant to solve wasn’t clearly defined in the first place.
So, the initial step is not about scanning the market for the most advanced tool or the most hyped technology. The first step is diagnosis— defining the problem clearly, validating it with data, and aligning it to business priorities. From there, the Four A’s help you stress-test whether a technology is the right fit.
In short, solving a pain point with clarity, letting data guide the choice, and validating against applicability, affordability, accessibility, and availability—that’s how you avoid the common pitfalls and build a tech stack that genuinely advances supply chain performance.
In your consulting experience, do you often see clients grappling with dilemmas where legacy systems, organizational culture, and the buzz around new technologies like AI come into conflict? How do you help them navigate that?
Neha Sorathia, Sr Principal, Accenture Strategy & Consulting: We encounter this often and it’s become even more pronounced with the current wave of AI .Many organizations feel compelled to jump on the bandwagon, either out of peer pressure—“our competitors are doing it, so we must too”—or simply to appear progressive. But the truth is, your competitor’s challenges are rarely identical to yours. That’s why we always steer the conversation back to first principles: What is the real business problem we are trying to solve? Once that anchor is established, technology becomes a means to an end rather than an end in itself.
The hard reality is that digital transformation—especially in supply chains—isn’t just a technology project; it’s a cultural and organizational shift. It can be disruptive and deeply demanding for the teams involved. That’s why we urge clients to step back and run a few critical “readiness checks” before committing to any technology investment:
Process readiness – Are the core processes mature enough to absorb and benefit from technology? If the processes are not standard technology only exasperates the challenges
Change readiness – Is the organization genuinely prepared to do things differently? Far too many programs fail not because the tools are flawed, but because people resist change. Adoption is the ultimate litmus test of success.
Data readiness – This is often the silent deal-breaker. Ambitious plans for demand sensing, predictive analytics, or AI collapse when the organization discovers it lacks clean, structured, and reliable data. Without a strong data foundation, even the most advanced tools are powerless.
These checks are deliberately tool-agnostic. They cut through the noise of buzzwords, forcing organizations to look inward before leaping outward. They don’t take much time, but they can save years of effort, cost, and frustration.
So, my advice is simple: don’t let the buzz dictate your digital strategy. Let your business problems, your process maturity, your people’s readiness, and your data foundations set the agenda. When those align, technology truly becomes a game-changer—rather than an expensive experiment.
How should organizations evaluate ROI for supply chain technology investments? Beyond the basic metrics, what should we measure, and how should ROI influence technology selection?
Rajat Sharma: A final point to underscore is that clarity of objectives is paramount. Establishing a well-defined ROI framework—encompassing financial savings, operational efficiency, compliance, risk mitigation, and capability building—ensures alignment across functions and positions technology investments to deliver measurable, meaningful, and sustainable impact. Without such clarity, even the most sophisticated technology risks failing to achieve its intended outcomes. In essence, ROI in supply chain technology is inherently multi-dimensional. It goes beyond immediate cost reductions or time savings to include adoption, process transformation, regulatory compliance, risk management, and the cultivation of a digitally proficient workforce. Organizations that define, measure, and monitor these dimensions from the outset are far better equipped to capture long-term, sustainable value from their technology initiatives.
Neha Singh: In the early days of technology adoption, the prevailing belief was that tools were “plug and play”. The assumption was that the moment you implement a system, results would appear immediately. From experience, we’ve learned that this is rarely the case. One of the most important steps is to establish a structured ramp-up period with robust feedback loops. ROI should not be measured purely on instant KPIs; it should account for capability building, user adoption, and integration into daily workflows.
A critical factor often overlooked is adoption metrics. Simply logging into a tool does not mean it’s being used effectively. Real adoption is reflected in the impact on the KPIs the tool was designed to improve. For example, if purchase orders were previously 80–90% manual and now 100% are generated through the system, that is true adoption. These metrics must be embedded in the tool itself, not tracked externally by analytics teams. Without this, your ROI calculations are misleading.
Tannistha Ganguly: I’d like to add a perspective that’s often overlooked when we talk about ROI: post-implementation enhancements and continuous learning. In my experience, after any software implementation, end users almost always come back requesting additional features or modifications. Some organizations see this as scope creep or a cost overrun—but that’s a narrow view. These requests are actually a sign that the tool is being used, understood, and valued. For example, in a recent WMS rollout in Brazil across two warehouse sites, the first site went live and, after a few months, the second site requested several enhancements based on learnings from the first. These weren’t arbitrary requests— they reflected a deeper understanding of the tool’s capabilities and how it could be applied more effectively to daily operations. While implementing enhancements incurs additional effort and cost, it also demonstrates adoption, engagement, and real-world utility—all key contributors to ROI.
Another important aspect is that these enhancements drive process improvements and operational efficiency. Users identify gaps, inefficiencies, or new opportunities that were not apparent at the initial implementation. By incorporating these enhancements, the organization captures additional value from the tool that wasn’t initially quantified in the original ROI calculation.
Additionally, there’s a capability-building dimension here. Each enhancement request reflects learning: users are becoming more digitally literate, discovering new functionalities, and expanding the tool’s usage beyond the original scope. Over time, this contributes to creating digital citizens— employees who can effectively leverage technology, experiment, and optimize processes independently.
So, when evaluating ROI, it’s essential to look beyond the first deployment. Adoption, enhancement requests, learning, and digital capability growth are all part of the return. Organizations that recognize and measure these elements can unlock far greater long-term value from their technology investments than those who focus solely on immediate financial metrics.
Shreyas Dhore: Traditional ROI metrics—cost savings, efficiency gains, and reduction in man-hours—are still important, but in supply chain, ROI is far more multidimensional. It’s not just about immediate financial gains; it also encompasses compliance, risk mitigation, operational resilience, and workforce transformation. Supply chains manage significant capital, interact with multiple external stakeholders, and operate under strict regulatory oversight. Even if a tool doesn’t reduce costs right away, if it ensures compliance, enforces procurement controls, or prevents operational errors, that is a meaningful ROI—avoiding fines, reputational damage, or downstream disruptions is a tangible return.
Equally critical is the concept of building “digital citizens” within the supply chain. Technology ROI should be evaluated not only on current outputs but also on its ability to upskill teams and embed digital capabilities across the organization. Are employees moving from manual, spreadsheet-heavy processes to using advanced planning tools, analytics dashboards, or AI-driven modules? Are they becoming proficient in leveraging data for decision-making rather than intuition alone? A tool that increases workforce digital literacy, empowers teams to act on insights, and encourages data-driven collaboration delivers strategic ROI that extends well beyond immediate cost savings.
In today’s context—where AI and automation are reshaping roles— investing in tools that enable digital citizens is especially critical. These employees become the backbone of a future-ready supply chain, capable of driving innovation, efficiency, and agility. ROI, therefore, is not only measured in dollars or hours saved, but also in the creation of a digitally capable, adaptable workforce that can sustain and scale organizational value in a rapidly evolving environment.
Neha Sorathia: Our approach to ROI begins by anchoring it to clearly defined, measurable KPIs. Before implementing any technology, we first identify the specific business outcomes we aim to influence—whether that’s service levels, inventory balance, supply plan accuracy, lead times, or overall operational throughput. These KPIs are not just metrics; they form the foundation of the entire technology journey, guiding design decisions, adoption strategies, and continuous improvement efforts.
A recurring challenge we observe is that organizations often focus on tool deployment rather than tangible outcomes. Users may log into the system, but if they continue relying on spreadsheets or manual workarounds, the tool has not truly been adopted. Logging in alone does not equate to effective usage. Without process integration, even the most sophisticated systems fail to deliver meaningful ROI. By linking ROI directly to KPIs, we create a concrete benchmark for success, allowing teams to measure real impact. For example, if the target for supply plan accuracy is 85% but current performance is 80%, the technology should be enabling teams to close that gap. If users revert to Excel or other manual processes, it signals obstacles—be it inadequate training, process misalignment, or system usability issues—that must be addressed immediately.
Another critical dimension is sustenance . Technology ROI is not just about initial performance; it’s about ensuring that the solution continues to generate value over time. Many implementations span 12–18 months, anchoring ROI to measurable outcomes ensures that sustainable usage is embedded from the outset, keeping teams accountable and aligned with long-term business goals.
What’s your perspective on adoption and utility of AI and ML? We know that not all organizations have implemented these technologies exactly as envisioned. How is adoption starting to take shape, and what should leaders keep in mind?
Neha Singh: When we look at the current market for AI tools, most solutions are very process-specific—for example, AI for demand planning, inventory optimization, or production scheduling. While these tools are powerful, there are two critical challenges that often get overlooked.
First is data readiness. AI, at its core, is advanced statistics at massive scale. It thrives on high-quality, structured data. Very few organizations—even industry leaders—can claim to have perfectly clean, comprehensive data for every process. Yet, most AI solutions assume that the data pipeline is flawless. In practice, significant manual effort is often required to clean, structure, and prepare data before it can be fed into AI models. This foundational step is crucial, but it is rarely built into the AI tool itself.
Second is algorithmic fine-tuning. AI tools are not plug-and-play. Even when you have a functioning solution, the underlying algorithms often need customization and calibration to match your unique business processes. This requires capable teams—people who understand both the process deeply and the statistical logic behind AI—to work closely with the solution providers. They translate business rules into algorithmic logic, tweak models, and ensure the AI tool aligns with real-world operations.
So, in summary, if an organization is planning to adopt AI in supply chain, it must invest in two foundational pillars: (1) clean, high-quality data, and (2) skilled teams who can bridge business processes and AI logic. Without these, even the most advanced AI tools will struggle to deliver meaningful results.
Shreyas Dhore: That’s a great question. Even before diving into AI and ML—the buzzwords dominating supply chain discussions—it’s helpful to reflect on how technology has shaped this field over the years. When I started my internship at Siemens in 2004, purchase orders were physically sent via courier. Over time, we moved to fax, then email, and eventually EDI. This evolution—from courier to fax, email, and EDI—is the silent backbone of today’s supply chains, enabling logistics, procurement, and warehouse operations.
An important point to note is that supply chain professionals have always been data analytics practitioners, even if informally. From day one, we’ve used spreadsheets, VLOOKUPs, HLOOKUPs, and basic statistical methods to make sense of operational data. This foundational skill set is what allows us to leverage AI and ML effectively today. AI and ML alone are not enough. Adoption and utility depend on:
Solving the right business problem first, not chasing technology for its own sake
Ensuring accessibility and availability for end users
Designing tools that complement existing skills and workflows
Applying a structured framework like the four A’s to guide implementation
When these elements are aligned, adoption becomes natural, and the transformative potential of AI and ML can be fully realized across the organization. Ultimately, technology must be a means to solve a core problem, not an end in itself.
Neha Sorathia: AI and ML have been part of supply chain conversations for some time, with demand forecasting being one of the earliest and most mature use cases. Over the years, organizations have refined these applications, improving both accuracy and usability.
Today, the focus is shifting toward agentic AI, where multiple specialized agents address individual supply chain challenges, coordinated by a master agent that integrates their outputs. This architecture is evolving quickly and moving from conceptual discussions to practical, real-world deployment.
There’s also a growing ambition to achieve autonomous, touchless supply chain planning. While fully autonomous planning is not yet a reality, organizations are experimenting with ways to make planning processes more self-directed and intelligent. The objective is to design AI agents that are highly aligned to specific business problems.
This marks a fundamental shift in AI application within supply chains. The technology is transitioning from isolated proof-of-concepts to integrated, scalable, problem-focused solutions. Given supply chain’s structured data, operational complexity, and high impact on business outcomes, it is proving to be a highly fertile ground for AI and ML innovation, enabling organizations to improve efficiency, enhance decision-making, and unlock measurable value across the enterprise.
(Disclaimer: The views and opinions expressed in this article are solely experts’ own and do not represent the official policy, position, or views of their employers or any organization with which they are affiliated. The content is based on their independent analysis, experience and expertise.)
Categories
