Most retailers know their last-mile systems are disconnected. Few recognize how much that fragmentation costs them in slower decision-making, higher cost to serve, and broken delivery promises.
On this episode of Deliver: The Last-Mile Performance Podcast, hosts Bringg CEO Guy Bloch and Sales Engineering Lead Raquel Zanoni sit with Retail Technology Executive and Transformation Leader Shweta Bhatia. Their conversation explores why siloed systems persist across enterprise retail and what it takes to connect them without starting from scratch.
Shweta brings more than two decades of retail technology leadership to the conversation. She led enterprise-wide digital transformations at Advance Auto Parts, Dollar General, Walmart International, and Kohl's, spanning thousands of stores, multiple distribution centers, and global technology portfolios.
Watch the full interview to learn:
- Why disconnected systems hurt performance but full consolidation isn't the answer
- Which last-mile KPI collapses first when systems don't communicate
- How leaders drive technology adoption by asking the right questions and putting people at the center of transformation
Get the key takeaways and full transcript below. Subscribe on Spotify, Apple Podcasts, or YouTube to get the latest episodes.
Where to listen and subscribe
Key takeaways
Siloed last-mile aren’t just a technology problem, they’re a business constraint
- Siloed systems are the natural outcome of organizations growing, scaling, and solving problems at different points in time. They rarely result from poor decisions.
- Complexity becomes a business constraint when simple questions become hard to answer, like understanding inventory across systems.
- The warning sign of too much complexity is when the business starts bending itself around the system: manual workarounds, tribal knowledge, and shadow tools become the norm just to keep operations running.
- Businesses too often rely on a handful of experienced associates to compensate for disconnected systems and it creates organizational fragility.
Retailers don't typically recognize the cost of fragmentation until it surfaces as slower decisions, higher cost to serve, and inconsistent customer experiences. At that point, the system is no longer reducing risk. It's creating it.
The first KPI to collapse when systems fail? On-time delivery.
- On-time, in-full (OTIF) is the first metric to fail when last-mile data is fragmented.
- Factors like order management, inventory visibility, and carrier execution can be solid individually, but the customer experience can fail because teams measure success locally and not collectively.
- An unexpected demand spike in a fragmented system, for example, can mean systems fail to share that constraint in real-time, which breaks the customer promise.
- If an organization cannot fulfill the customer promise, the failure is organizational and not limited to any single team or system.
OTIF is the clearest signal of whether last-mile systems are truly integrated or just technically connected. When that metric slips, the root cause is almost always a coordination gap, not a capacity gap.
System consolidation and modernization shouldn’t be the main goals. Adaptability and outcomes are.
- Complete system consolidation is not realistic for most organizations and should not be the goal. Systems exist for good reasons and were built to handle scale, complexity, and real operational constraints.
- Smarter coordination is the solution: let systems do what they do well while creating a shared real-time layer that aligns decisions across all of them.
- New technology investments should answer one question clearly: "What decisions will this system improve?" If a solution cannot articulate improved decision-making under real-world constraints, it adds complexity, not value.
- A system that truly improves last-mile performance provides clarity, reduces handoffs between teams, and helps the organization respond more intelligently when conditions change.
Modernization is not about replacing old systems or consolidating platforms. It's about restoring flexibility so the business can grow, adapt, and keep the customer promise intact.
Last-mile technology adoption fails without people at the center.
- People account for the largest share of any transformation. Technology represents roughly 20-30% of the effort and process another 20-30%. The remainder is people.
- The hardest leadership challenge is rarely the technology itself. It's the uncertainty that new systems create for the people who use them every day.
- Resistance manifests as workarounds, quiet disengagement, or silence in dialogue (a major red flag during transformation).
- Adoption accelerates when leaders make the people using the new tools and processes feel supported, confident, and part of the change.
At scale, transformation only sticks when people believe the change is happening with them, not to them. The organizations that build internal champions early and lead with transparency move faster and sustain results longer.
The future of the last-mile isn’t speed, it’s adaptive performance.
- The industry is shifting from step-by-step optimization to real-time intelligence: sensing what's happening and adjusting before customers or teams feel the impact.
- New technology will support better decision-making for people on the ground, helping teams understand trade-offs, respond to constraints, and act with confidence when conditions change.
- Adaptive—not just reactive—last-mile logistics improves cost, speed, and builds trust for both customers and operational teams.
- Leaders should treat the last mile as a trust problem to solve, not a cost problem to manage. When systems, data, and decisions align around that promise, speed, cost, and scale follow.
The retailers and logistics providers that design their last-mile operations around trust will outperform those still chasing incremental efficiency gains across disconnected systems.
The interview: Why fragmented systems fail enterprise retailers
Guy: Shweta, over the years you've led to technology transformation at the largest retailers out there [like] Walmart and Kohl's and the Dollar General and recently with Advanced Auto Parts. But I want you to go back to the beginning of your career and try to remember the first time the last mile came to you as a problem; when you looked at it and said, “This is broken and I can fix it. I actually want to fix it.”
Shweta: It takes me back 15 to 17 years ago. Early on in my career, what really stuck with me wasn't a single broken tool. It was a broken assumption. And if you go back, we designed last-mile systems as if customers, stores, last-mile couriers, and associates all lived in a very perfect, neat, and predictable world with their own lanes.
And in reality, the last mile is messy. It's human because it's centered around human beings and it's constantly changing. I remember watching stores get blamed for missed deliveries, carriers getting penalized for constraints they did not control, and customers being left completely in the dark. While the technology stitched all of this together, it was after the fact; it was not in real time.
That really stuck with me. And the moment it clicked for me was—the systems weren’t designed to orchestrate outcomes, they were designed to record failures. And over the course of my career that insight stayed with me. The opportunity isn't just about faster delivery. It's building platforms that adapt in real time and respect and honor operational realities on the ground, because that's where the magic happens—they center [around the] human experience on both sides of the transaction. So that's truly what last-mile transformation means to me, moving from static execution to intelligent orchestration.
Guy: I'm thinking of how hard it is when your systems are completely fragmented and you're trying to go after an outcome, but that outcome is dependent on so many different systems that don't talk to each other. Over the last few years, [there have been] major efforts by every retailer, every CEO, every leader, to transform their tech stack and make it less fragmented and more integrated so they can orchestrate better. Yet, when you look at many retailers, they're still fragmented. Why is that? Why is it so hard?
Shweta: There may be multiple systems but they don’t necessarily talk to each other. I've seen this pattern repeat across many large organizations and it's rarely because anyone made poor decisions. Siloed systems are usually the natural outcome of organizations growing, scaling, and solving problems at different points in time. So it builds over a period of time.
And in retail, [there are many functions]: a commerce function, a supply chain function, store operations, and last-mile functions. The systems are built to support outcomes in the specific lane they’re accountable for. Those systems often work well within their specific domains, but end-to-end coordination was not always the primary design objective at that time.
Taking us back 10 years, retailers optimized and designed for their own P&L (profit and loss). [They made] good decisions for that domain, but they weren’t integrated. And every tech leader has had digital transformation on their mind for the last five to 10 years. It has become more incremental than architectural. We've added tools, we integrated for visibility, and we worked around constraints, which made perfect sense in the moment, rather than stepping back to redefine the experience we ultimately wanted to enable, both from a customer perspective, and from an associate perspective. That's why systems can technically talk, but they struggle to work together when it matters the most.
When system complexity becomes a business constraint
Guy: From your experience leading large scale technology transformation projects in really large companies, how can companies identify when system complexity starts holding them back and not giving them the impact they want?
Shweta: You usually know complexity is holding you back when simple questions become hard to answer. If it takes multiple systems, spreadsheets, and caveats to commit to a delivery date or understand what your inventory position is, the system is already working against you, and you just have to see through it.
Another signal is when workarounds become normal. You have manual steps, tribal knowledge, shadow tools just to keep things going. And finally, decision-making slows because leaders don't have a shared real-time view of reality. At that point, complexity isn't an IT issue anymore. It's a business constraint, not a technology constraint.
The biggest one for me is when it becomes a business constraint and when it inhibits the business to move forward in an unrestricted way. That's when we need to solve for it.
Raquel Zanoni: Can you give us a real example? When did you have to keep a legacy system alive? And what was the moment you realized, “Okay, now this is holding us back.” What really pushed that decision?
Shweta: I have kept legacy systems alive more times than I can count. Often for very good reasons. They were stable. They were deeply embedded in operations and replacing them overnight would have created more risk than value because we have to be very pragmatic of the world we live in. The moment I knew a system was holding us back wasn't when it was old. It was when the business started bending itself around the system.
In one case, a core execution platform couldn't support real time changes to delivery promises. Teams compensated by adding manual steps, building exception processes, and relying on experienced, but a very small handful of associates, to make it work. At first, it felt like resilience and it was. But over a period of time, it became fragility.
And what pushed the decision wasn't a single outage. It was a cumulative cost, slower decisions, higher cost to serve, inconsistent customer experience, and increasing dependency on just a handful of resources. That's when it became very clear that keeping the system alive was no longer risk mitigation—it was a risk creation. We were increasing risk for the business. And that's the time when you act.
So for me, modernization wasn't, in that case, replacing old systems. It was about restoring flexibility to enable and grow the business.
Guy: Is complete system consolidation realistic for most logistic organizations or is it more around the goal of orchestrating among different systems?
Shweta: I don't think complete system consolidation is realistic for most organizations. And I don't think it should be a goal. In large logistics environments, systems exist for good reasons. They have been built over time to handle scale, complexity, and very real operational constraints. Forcing everything into a single platform probably creates more disruption than progress, given the period we are talking about.
What's far more effective is smarter orchestration, letting systems do what they're good at while creating a shared real-time layer that aligns decisions across all of them. When that's done well, leaders don't have to choose between stability and agility. You preserve what works.
Reduce unnecessary complexity, and give the organization ability to respond as conditions change and as your business changes because we live in a world where every day there is a new NVIDIA chip that's being launched. So that's why these systems and these platforms need to provide that agility to the organization. For me, that balance becomes very important. Continuity with adaptability is where real transformation happens.
The first KPI to collapse in a fragmented system
Raquel Zanoni: You mentioned some negative outcomes of not having those systems talking to one another. During times of disruption, that's when we feel it the most. Is there an example from your experience of one outcome or one KPI that you think collapses when systems aren't talking to one another?
Shweta: The first KPI that collapses when last-mile data is fragmented across different systems is on-time, in-full delivery because it depends on systems actually working together in that moment.
I've seen situations where order management, inventory visibility, carrier execution, and store operations were all technically green in their own systems. But the customer experience was failing and each team was measuring success locally while the outcome that mattered was slipping. So the KPI that matters the most is on-time, in-full delivery. If we cannot fill that customer promise, we are failing as an organization.
Raquel Zanoni: Do you think that on-time delivery is what retailers should be looking at when they're wondering if their systems are siloed? What is the threshold for good or bad in terms of that particular KPI? And what should a retailer see and think to themselves, “Okay, we might need to fix this?” Is it truly individual or is there a standard across retailers?
Shweta: I'll give you a very simple example of it failing from a customer perspective. In one case years ago, the demand spiked unexpectedly just as carrier capacity tightened. Inventory looked available, orders were released, stores were staffed. So all these different lanes were perfect but carrier capacity was a little tightened and the last-mile execution couldn't adjust because the systems were not sharing that constraint in real time.
So all those other lanes were working, but then this one lane had a constraint and it was not sharing. So customers received inaccurate delivery promises. Stores absorbed the fallout…they get penalized for missing the deliveries and the cost of service. Now think about that from a CFO perspective, the P&L perspective, and the cost to serve from a supply chain perspective, and finance perspective spiked up.
How to evaluate new technology investments
Guy: As I'm hearing you and from the perspective of a technology leader, what you're saying is that first of all, the world is fragmented. We are surrounded by a fragmented supply chain and the last mile is dependent on that. There is always the aspiration to consolidate and bring systems together with more alignment so you can better orchestrate that.
But as a CEO that sits between those two dimensions, what is the right thing to do, in your mind? Is it really to consolidate or is it orchestrating among different systems and microservices in the last-mile?
Shweta: It is never about consolidation to me. Technology consolidation is a far-fetched dream for a lot of leaders. That's not reality. We need to focus on the outcome and the experiences we want to enable for our customers. Modernization isn't really about consolidating and doing more. It's about doing a few of the correct things well. There are budget constraints and competing priorities. There are constraints all across [the last-mile]. That's the reality of the world we live in. With that lens, consolidation never becomes a priority.
It’s more about orchestrating what you have and enabling and adding more capabilities that you don't have to make sure that that fabric of the organization is anchored towards the customer outcome. What experience do you want the customer to have? And how do you want your store associate experience to enable the customer [delivery] experience? The customer [delivery] experience will never be better than the store experience because your stores actually enable the customer experience.
Guy: So it’s not just bringing systems together and orchestrating among them. You're also looking at new technologies, new capabilities to bring into the organization to keep it up to speed and on par with what is available.
From that perspective, [how do you know] what's going to improve last-mile performance versus just adding another layer to something that already exists? When you're measuring another capability or functionality, another area of the last mile, at what point do you know it's going to actually move the needle versus being a nice-to-have, but it doesn’t deliver the business outcomes I want?
Shweta: I start by asking a very simple question: “What decisions will this system make?” Or, “What are we trying to do where this system will make it better, faster, or act with more confidence?” If a solution cannot clearly articulate any improved decision making in the last mile, especially under the real world constraints, it's likely to add more complexity, not value. So I start with a very simple question. Because at times, the answers are complex but the questions are simple.
I also look at whether the system fits into the organization's reality. Can it adapt in real time? Does it integrate naturally into the existing workflows? Does it make life easier for the people actually doing the work? That goes back to the folks in the stores, the folks in the distribution centers and the associates who actually are serving the customers, which is very, very important.
And finally, I pay very close attention to data—not just whether it exists, because there is an insurmountable amount of data that we are surrounded by, but whether it's trusted, shared, and actionable across teams. When a system provides clarity, reduces handoffs between different teams and helps the organization change more intelligently, that's when you know it's truly improving your last-mile performance.
Raquel Zanoni: You mentioned three outcomes to pay attention to. They were: making things better, faster, and giving people more confidence in what they do, or making it easier for them to do their work. Can you give us an example of what that means?
Shweta: A system that truly makes things better adjusts delivery outcomes before the customer ever sees it. A system that makes things faster allows teams to read out or rebalance capacity without manual trade offs. And a system that builds confidence gives the store team and customer service a shared real-time view of what's actually possible so they're not guessing or apologizing after the fact and the call center is getting calls.
When people trust what a system is telling them, they stop working around it and start working with it. That’s when performance improves—not because people are working harder, but because people finally believe in the system that supports good decisions.
When to build versus buy
Raquel Zanoni: I like that idea of trusting the system or finding a system that your associates, whether they're in the store or operationally on the backend, can trust. To pivot a little, one thing that we see when it comes to trusting a system and finding the right one is that many organizations don't know where to start. They have a lot of internal dialogues like, “Should we go out to the market and find a system? Should we build something ourselves?”
How often have you been in that decision-making scenario of build versus buy? And was there ever a time that you thought, “I'd go back and make a different decision?”
Shweta: It's a very classic retail and technology leadership problem. And I have been in that debate many, many times. It's rarely about capability alone—it's about the focus.
In one situation, the team was split because we could build the solution and there was a lot of pride in that. When you have engineers and good product people, they want to build. But when we stepped back, the real question became, “Is this where we want to differentiate or is this where we need reliability and speed?”
What tipped the scale was realizing that building would tie up critical talent because everyone has a limited pool of talent. Buying allowed us to move faster in that scenario and keep our teams focused on what truly made us unique.
Looking back, I would probably make the same decision but pair it with stronger change management and clearer ownership from day one. Build versus buy isn't a one-time choice. It's an operational commitment; it's an organizational commitment. Success depends on how much you adopt and what you choose. It's a combination of the choices you make as an organization: build versus buy versus co-build versus partner. There are a lot of choices you make as an organization that [eventually] become your operating routine.
Raquel Zanoni: What happens to organizations that make the wrong choice? Have you seen that happen? And at what point did they ask whether they made the wrong choice?
Shweta: It has definitely happened to me. It's a common thing. There are no mistakes. You just learn from those learnings. And those learnings are what we, as the technology leaders and teams, reflect on.
It took us two years to build an inventory system. Inventory is table stakes in today's world. As an organization, do you have the capacity? Is that a priority? Is that what you want to be differentiated with? Or do you want to take a platform and do intelligent integration, orchestration, and have real-time inventory across your different systems? It's a choice every organization makes. I don't know if it's a good or a bad choice.
There are things [where organizations say], “We'll never build an ERP.” There are organizations that [could build one really well]. Every organization might choose a little differently based on their strategy, their operating principles, and whether they can support it. Can they put resources in it, not just today, but in the future to support and maintain that system?
Last-mile technology adoption fails without people
Guy: And we know that it's not just about technology—it's about people and adoption and scaling an entire organization. In your experience, what were the hardest leadership challenges that you had to deal with and how did you address them?
Shweta: To me, the hardest leadership challenge is rarely about technology—it is the adoption of the technology, no matter how smart and how good that technology is. It is the uncertainty that it creates for people. The biggest leadership challenge I've faced personally is trust.
When systems change, people worry about losing control. They worry about losing their ability to do their jobs well. And if leaders don't acknowledge that openly, resistance shows up in different ways, subtly at times and in spoken ways, or with workarounds or quiet disengagement.
Silence is never a good thing. I always believe that when you are open as an organization and talking about the problems and the challenges in a very constructive way, that's when true transformation happens because people are the asset to any organization. What's worked for me is leading with transparency and respect—being clear about why we are changing, what will be different, and importantly, what won't change and what won't be different. What we will do is very important, but what we are not going to do is equally important as well.
I spent a lot of time listening to teams closest to that work and incorporating their feedback early on, not after the decisions are logged. I also learned that adoption accelerates when you invest in capability, not just training. When people feel supported, confident, and part of the situation, they move from being compliant and cautious of the system to [feeling like], “I own the system and I own the outcome of the system.” And at scale, transformation only sticks when people believe the change is happening with them, not to them.
Raquel Zanoni: You mentioned resistance and how that shows up in big and small ways. Do you have an example of what you mean by that?
Shweta: There are three elements that are a part of any large transformation: people, process, and technology. To me, “people” is the most important part of any transformation. Technology is 20-30% and process is 20-30%.
I fortunately, and unfortunately, have a lot of experience where, if done right, people become those champions; they become the advocates that redesign the entire workflow and believe in the outcome of where this new transformation could lead us. If they are not involved from the beginning, or we don't build and identify those champions and bring them along in the journey, then people become silent. And when you get a lot of silence in the room, there is a problem, because silence is never good in transformation rooms.
So I always encourage healthy, open dialogues. We can agree to disagree, but let's talk. Silence is never a good thing. And when you get that, then there is definitely a problem.
The future of last-mile performance
Guy: What are you most excited about in the last mile today? And what do you think is on the horizon for the next couple of years?
Shweta: What really excites me about the last mile right now is the shift from optimization to intelligence. For a very long time, we focused on making individual steps faster and cheaper. What's changing is our ability to sense what's happening in real time and adjust before customers or teams feel the impact. And that's going to be a big shift.
[The thing] on the horizon and what I'm most encouraged by is technology finally supporting better decision-making for people on the ground; systems that don't just automate tasks but help teams understand trade-offs, respond to constraints, and act with confidence when conditions change. When that happens, the last mile stops being reactive—it becomes more adaptive. That's when you see improvements not just in cost, but in speed and most importantly, in trust for customers and for the teams that are doing the work every single day.
Guy: What a complex world. In the last-mile, we make a promise to a customer, then we deliver on that promise. That's what the customer sees—two touch points. But behind that, it’s such a complex world with different systems and fragmented environments we must bring together.
We have one more question. You have so much experience and we know retailers and logistics players all understand that the last mile is fundamental to compete in the world of eCommerce. Everyone is stepping their game up and setting plans for the year. What advice do you have for leaders and should they focus on over the next couple of years?
Shweta: The last mile is where rubber meets the road, in every sense, in today's world. Don't treat the last mile as a cost problem to manage, treat it as a trust problem to solve. Every delivery promise is a moment of truth for your customer and for your teams. When systems, data, and decisions aren't aligned, that trust erodes very quickly and very quietly and you might not know until it's too late.
The leaders who get this right focus less on optimizing individual steps and more on orchestrating the entire experience so the organization can adapt in real time to keep the customer promise intact. When you design around the last mile around trust, everything else—speed, cost and scale—follows.
