Karthik Naig made his first foray into warehouse robots at Greyorange in 2015. After helping build the software for Butler robots and scaling the business at Greyorange, he was excited with the cloud approach to robotics, and has been driving the PA AMRs initiative at Rapyuta Robotics since.
Hailing from Bangalore, he is currently living in Tokyo with his wife and daughter. If you ever happened to drive by Sano on a lazy Saturday afternoon, you may just happen to see him playing cricket with the rest of his team. His other interests include spirituality and meditation, writing and watching movies.
Good to finally get hold of you Karthik from your busy schedule. What’s been keeping you so much occupied lately?
Mostly I am focused on successful operationalization of our PA AMR (Pick Assist Autonomous Mobile Robots) product as well as helping build the next set of features to make it more compelling to the end user. We are also looking at technology improvements, especially in perception and navigation, so that we can help warehouses with more diverse layouts.
Let’s go back a bit. What’s been your experience with warehouse robotics?
My first experience with warehouse robotics was with GreyOrange robotics in India. I was the software product manager for the Butler robot – a Goods-to-person (GTP) warehouse robot. When I joined, I knew neither robotics nor warehousing. But the learning was fast since we worked very closely with Flipkart and other retail and eCommerce customers, and we successfully drove its adoption across geographies.
For the last couple of years, I have been at Rapyuta robotics in Tokyo, building the hardware and software for the PA AMR robots and driving its adoption in Japan. Right now, we are in the process of going live at several warehouses in Tokyo.
What are the main barriers to warehouse robotics adoption?
The most basic factor is awareness. While a lot more people are now aware of warehouse robots than was the case, say 3 to 4 years ago, there is still a long way to go in terms of knowing exactly how the robots are used and what is to be expected from them. As a result, significant effort is required in educating customers about the use cases and business value.
Awareness is key in Japan for another very important reason – decision-making – and that is part of the culture. Unlike other countries, decision-making in Japan is a collaborative process, not driven top-down. So while business owners may want to deploy robots, warehouse managers play an important role in decision making. And winning the trust of warehouse managers takes time. So trust building at various levels is key to successful adoption of robots in Japan.
The second barrier is reliability. This is a broad generalization, but I think warehousing robotics has still not managed to reach the levels of reliability at scale that is expected from the older automation systems. It is seen as “cool” but warehouse managers don’t want “cool”; they want “predictable operations”.
The third barrier is cost. This is true especially of GTP systems which require high capital cost and large changes to their existing warehouses. Although Amazon has adopted robots at scale, most 3PLs and ecommerce customers are wary about making high capex investments.
For reliability, Rapyuta addresses the problem by using a multi-layered approach with a core robotics platform, a robotics applications layer and finally the warehousing applications layer on top. This isolates the problems to each layer making the solution more stable. Also, it makes development very easy, hence we have made rapid progress in a short duration.
For minimizing costs, Rapyuta offers robots-as-a-service which minimizes the capex expense. RAAS enables customers to adopt robots for retro-fitting warehouses and scale up and down as needed.
What are the challenges unique to AMR adoption in warehouses in Japan?
Japanese warehouse managers place very high importance on customer satisfaction. Naturally, they are apprehensive about technologies which have not yet seen adoption on a mass scale. Further, delivering high throughput is more of a challenge in Japan since most warehouses are small and have narrow aisles which limits the number of robots and their speeds.
So the adoption is a little more gradual right now but I expect that to change soon.
Do you expect that there will be a domino effect in robotics which can drive mass adoption in warehouses?
Yes, of that I have no doubt. On the one hand, the eCommerce revolution has definitely swung the pendulum towards small orders, leaving robots as the only realistic technology option for picking processes. Labor costs are not getting any lower, and real estate for warehouses are getting more expensive as warehouses move towards the city. So warehousing costs are going to put massive pressure on profit margins.
Also, robot performance and safety has rapidly improved and the hardware has become cheaper and more flexible. I would say that we are at that phase where the true expectations from AMR capabilities are being realized. Once a couple of major 3PLs or eCommerce customers adopt AMRs on a large scale apart from Amazon, there is no option for the others but to follow.
The thing about the Japanese, as is universally known, is that they are the pioneers in automation. But the adoption tends to be a little gradual until the technology is proven. This arises from their passion for customer satisfaction. Once they are convinced about a technology, the domino effect is a natural consequence.
How do you see warehouse robotics evolving in Japan? Will it mirror the rest of the world?
I see a few major steps in the global evolution of warehouse robotics. The ultimate goal I can see is the concept of dark warehouses, where every operation is done by robots and there are no humans at all. Now that’s a bit like fully autonomous cars. Sure we know we will get there in the long term, but there are many steps in between which offer a lot of opportunities for disruption.
In the medium term, I see the combination of robotics with warehouse-as-a-service providers as a major offering. Take grocery, for instance. World over, grocery warehouses deal with the same inventory, similar storage requirements and similar transportation pressures. So they need to optimize for speed without sacrificing flexibility. I think this is where the combination of the older, more rigid technologies with the newer AMRs can lead to build warehouse-as-a-service.
As for Japan, I think the more interesting question is whether it will lag or lead the world. Although Japan may be relatively late in adoption of warehouse AMRs, we should note that there is a tremendous interest and appetite to use AMRs. I believe it will soon be a leader in AMR adoption as the technology matures.
How is Rapyuta different from other companies in the warehouse AMR space?
I think the primary difference is that Rapyuta brings a very fresh philosophy to the market.
Rapyuta believes in a platform driven play. By focusing on our core competency, we leave the field wide open for business application software and Integration with third party software. Even for hardware, we believe in partnering with robot makers for anything other than the one robot type we have developed in-house. Our focus is very clear – to give the customer the best solution possible. For that, we stick to doing what we are best at – which is the core robotics platform and software – and partner for the rest of the solution.
What are the challenges and opportunities unique to the warehouse PA AMR space?
I think the main opportunity is to make AMRs a generic enough offering that can be deployed in virtually any warehouse. The challenge here is the technology, mainly the navigation, which needs to be improved to a level that robots can travel in any environment. In my view, this will be through a combination of better sensors and better software so that inputs from multiple sensors can be fused together to make the right decisions.
The second opportunity is to make installations nearly instantaneous. The main bottleneck here is that most warehouses have legacy WMSes or ERPs which are usually heavily customized. As a result, integrations take up several months unless there are adapters for the specific WMS. I think the most promising solution in this space is Warehouse Execution Systems (WES). Rapyuta is open to working with promising WESes who can help reduce the integration workload without sacrificing flexibility.
The third is to enable collaboration between different types of robots. This requires an understanding of the main inventory flows in the warehouse and automating them in a repeatable and efficient way. This is where I see rapyuta.io, Raypyuta’s cloud robotics platform in conjunction with the WES, as holding tremendous promise for solving the challenge for both scale and scope.
Overall, I think warehousing robotics is a very exciting space to be in at this time as it’s poised for takeoff over the next few years.
The PA-AMR solution built by Karthik’s team fast forwards your solution development with production ready infrastructure, fleet management AI engine, hardware and software catalogue, and simulation for multi robot operations in dynamic environments. Check it out here.