Autonomous mobile robots are becoming more common in factories, warehouses, and other industrial environments, but moving from a controlled pilot to dependable production use requires more than adding robots to an existing workflow. Integration with legacy systems, changing layouts, human workers, and multi-vendor equipment can determine whether a deployment scales or stalls.
In this TNGlobal Q&A, Rahul Nambiar, CEO and Co-Founder of Botsync, discusses the operational realities of deploying AMRs, the importance of multi-vendor orchestration, the metrics companies should use to judge automation projects, and what Southeast Asian deep-tech companies need to build credibility as they expand globally.
Autonomous mobile robots have moved from an emerging concept to a more visible part of industrial automation. Where do you see the biggest gap today between an impressive pilot and a deployment that delivers dependable value in production?
The gap usually isn’t in the robot itself. It’s in everything around the robot. A pilot only has to prove that automation can work in one corner of a facility under controlled conditions. Production deployment has to survive changing layouts, fluctuating volumes, and integration with existing systems day after day.
There are two gaps we see most consistently. The first is the integration layer. Most facilities run a patchwork of machines and software from different vendors that were never designed to communicate with each other. Connecting a new AMR deployment meaningfully to existing WMS, ERP, and conveyor systems – that’s where projects stall, not in the robots themselves.
The second gap is adaptability. A pilot environment is static by design; a live facility is not. Layouts shift with seasonal inventory. New equipment gets introduced. Workers take informal shortcuts that no one documents. Our MAG AMRs use a deep-learning-based navigation approach that’s built to handle these dynamics, so sites can go live within days and adapt as the facility evolves rather than breaking down every time something shifts.
The real test of dependable value isn’t the demo. It’s whether the system still holds up after the layout changes for the third time that quarter, and whether the operations team can manage it without waiting for a specialist to fly in.
What operational problems are companies in FMCG, food and beverage, and automotive most commonly trying to solve with AMRs, and which outcomes should they measure before deciding that a deployment has succeeded?
Across these sectors, the demand is being driven by the region’s growing need for scalable automation solutions, and it typically comes down to three recurring pressures, including labor shortages that make it hard to sustain consistent throughput, the need for business continuity when manual processes are vulnerable to disruption, and pressure to lift productivity.
Deployment of AMRs should aim at alleviating these pressures. Therefore, in weighing outcomes, companies must look at these four areas: Labor cost savings from reallocating headcount, quality improvement from more consistent execution, incremental production volume from optimized throughput, and operator safety in higher-risk environments. One thing to note is that these benefits must be weighed against the total investment, which includes not just hardware and software but also system integration, training, maintenance, and long-term scalability.
The outcomes also need to be mapped before deployment begins, not after. Our approach is to run onsite time studies and detailed workflow simulations before a single robot is deployed, so the ROI case is grounded in the customer’s actual operational data rather than a generic promise. Our Thailand deployment for a global automaker is a good example of this. After careful evaluation and onsite simulation, we deployed a system that completed over 40,000 production trips in 15 months, delivered over 1.2 million stamping parts, enabled reallocation of a forklift and two operators, and achieved ROI in under two years.
Why do real-world deployments often become more difficult once robots are introduced into active facilities with changing layouts, variable inventory flows, and human workers?
A pilot environment is static by design, whereas a live facility is not. Layouts shift with seasonal SKUs, inventory flows change with demand, and you have human workers moving through the same space as the robots.
Automation isn’t without its headaches, and implementing a new system can feel like assembling a puzzle where the pieces don’t come from the same set, because most facilities are running a mix of manual processes and automation simultaneously, not a clean automated-only environment. That’s why we think of AMRs less as replacements and more as co-pilots to the existing operation: machines that plug into an environment that’s already in motion, working alongside people and manual processes rather than requiring the facility to reorganize around them.
That coexistence is the point: since most companies cannot afford a complete transition to a fully automated facility, the automation has to be collaborative by design, adapting to the humans and processes already there rather than asking the facility to adapt to it.
Many operators have accumulated equipment from different robotics, software, and automation vendors. Why is multi-vendor orchestration becoming more important, and what does meaningful interoperability look like in practice?
Most operators didn’t set out to build a multi-vendor environment. It happened because they bought AMRs, conveyors, WMS software, and automation equipment from different vendors over time, each chosen for its merits, each with its own control logic and data format. The problem is that this results in a facility where machines are working toward the same goal but can’t talk to each other, and production data ends up trapped in silos where it can’t be acted on.
Meaningful interoperability isn’t just getting machines to communicate, it’s about what becomes possible once they do. When every machine in a facility feeds into one centralized platform, you can collect production data from every stage of operations simultaneously, spot bottlenecks across the whole system rather than within individual subsystems, and use AI to surface insights and take actions that no single-vendor view would ever reveal.
In practice, this means connecting AMRs, AGVs, robotic arms, conveyors, PLCs, and business systems – regardless of who made them – through one common interface. This shifts the conversation from “how do we integrate another robot” to “how do we use the data we’re now collecting to make better decisions.” The real value of interoperability is not just coordination, it’s intelligence. This is exactly what we built SyncOS™ to do. We strongly believe the next frontier of robotics won’t be in just making smarter individual robots, but in building intelligent systems that can orchestrate different robots together more effectively.
What are the most common technical or organizational obstacles that prevent AMR programs from scaling beyond one site or one narrow use case?
The obstacles are as much organizational as technical. Technically, it’s usually the integration burden, where every new site, every new vendor system, every layout variation reintroduces the interoperability problem if the underlying platform isn’t designed to abstract that away.
Orchestration is a big part of this: without a layer that sits above individual robot fleets, every new vendor or piece of equipment becomes its own integration project instead of a repeatable pattern, and that’s what usually caps a program at one site. Organizationally, it’s the reliance on specialist technical resources: if scaling to a new site or use case always requires deep engineering involvement, most operators simply don’t have the bandwidth to do that repeatedly.
When evaluating automation projects, what should companies consider beyond headline measures such as throughput or labor savings? Are there other indicators that reveal whether the system is improving operational resilience?
Throughput and labor savings give us insights that the system is working. To understand a system’s resilience, however, I’d also look at: how the system performs when conditions deviate from the norm (a demand spike, a layout change, a vendor outage), how quickly a facility can onboard a new use case without a fresh integration cycle, and whether the data the system generates is actually being used to improve operations, not just log them.
I’d also check whether the ROI case still holds up once the system is live, not just at sign-off. The strongest programs validate results with a single unit first, then scale deployment gradually, rather than betting everything on one big rollout. A more resilient program is one that started small, proved itself, and scaled gradually, so absorbing a disruption or moving to a new site doesn’t require starting over.
Botsync has expanded from Southeast Asia into the U.S. through strategic partnerships. What lessons has this transition offered about competing in a more mature automation market, particularly for a company that began in the region?
For markets like the US, Australia, and South Africa, our approach is to expand through existing customers or trusted partners who understand the local market, rather than opening cold. That approach is slower on paper but significantly more durable in practice.
What we’ve learned is that mature markets demand evidence at every layer. Companies are looking for proof, track record, and clear differentiation, not pitch decks. The credibility we’ve built through deployments with enterprises like Ford, Coca-Cola, Caterpillar, Kimberly-Clark, and Nestlé travels, and that matters enormously when you’re entering a market where established players have decades of relationships.
There’s also something that’s surprised us positively: the resilience we built by deploying in dynamic, resource-constrained environments across Southeast Asia and India turns out to be exactly what mature markets are now looking for too. In 2025, Botsync recorded 240% growth in production trips, crossing one million live production trips, and 230% revenue growth year-on-year. This was driven primarily by expansion from existing customers rather than new logos. That’s the signal that matters most to a new market: that the customers who know us best are the ones growing fastest with us.
Hardware and deep-tech companies often face a different funding environment from software startups. What does it take for Southeast Asian industrial technology companies to build credibility with customers, partners, and investors as they pursue global expansion?
The funding dynamic for hardware and deep tech is real. The capital requirements are higher, the timelines are longer, and the risk profile looks different on a spreadsheet than a pure software play. But I’d push back slightly on the idea that this is a fundamental disadvantage. What it actually demands is discipline: you have to build credibility through evidence at every layer rather than momentum and narrative.
With customers, it’s deployed systems and measurable ROI, not projections. Meanwhile, with partners, it’s a track record of integrations that actually hold up in production. With investors, it’s showing capital efficiency and sustainable growth alongside the top-line numbers. Not just scaling fast, but scaling profitably.
Editor’s note: This Q&A has been lightly edited for clarity and style. The responses remain those of the interviewee.
Share your perspective: TNGlobal welcomes contributed insights and expert commentary from across Asia’s technology and innovation ecosystem. Submit a contribution for editorial consideration, or explore more conversations in our TNGlobal Q&A and Interviews archive.
Singapore’s Robotics startup Botsync secures additional Series A funding from SGInnovate