The rapid advancement of Physical AI technology has brought significant attention to robotics startups. They are showcasing their technology through novel and often spectacular demonstration videos—robots folding laundry, washing dishes, or even performing martial arts and marathons.
The level of public interest is immense, attracting astronomical amounts of investment, and expectations for robots are higher than ever. Based on the current buzz, it might seem that robots built on Physical AI are on the verge of commercialization and that a large-scale deployment in industrial environments is imminent.
But when considering the proper and widespread adoption of robots in the industrial sector, do people truly grasp that the current state is merely the tip of the iceberg?
It doesn't take long to create a prototype and demonstrate a compelling technology demo. When we first started our company, it took us less than a month to purchase a research mobile robot, develop the necessary software, and conduct our first successful demonstration. And this is precisely the stage where the majority of robotics companies and their flashy demos reside.
The next phase is slightly more challenging. It involves:
This is where the development of a prototype for an industrial-grade robot form factor begins. The developed hardware prototype is then integrated with software to make the robot operational. Once ready, a Proof of Concept (PoC) test for the targeted task is conducted at the client's site. After several months, a single task is successfully automated in the customer's environment.
This marks the "demonstration phase"—showing that the task can be done. This is what we call Lab-Level product development.
At the Lab-level, one success counts as success. For Industrial-grade, however, one failure counts as failure.
The real journey begins now. To deploy even a single robot into a client's production line, it must operate continuously and flawlessly in the industrial environment. For a demo, successfully completing the task once, after perhaps 10 or 100 attempts, is often sufficient. Producing a video of that successful attempt and showcasing the technology is an easy win.
Operating flawlessly in an industrial setting requires the exact opposite: 100% success—whether it's 1,000 or 10,000 operations. This demands:
To develop such a complex system that operates without error, a systematic product development methodology at the System Engineering level is essential.
The truly difficult part follows the successful development. A product's existence is only justified when its clear utility value to the customer is proven. Once the robot is fully validated in a pilot line, its utility must be verified through the deployment of multiple units at that site.
First, the robot must pass the crucial hurdle of stakeholder acceptance at the work site. A natural resistance often arises because existing work processes are being changed. This involves convincing various parties:
Furthermore, seamless integration with the upstream and downstream processes is critical to ensure production remains unaffected. If the robot causes disruption to existing workflows, the difficult situation of having to remove the robot may become unavoidable.
The most decisive factor is productivity. It's not enough to confirm the robot's task performance; the overall production throughput must be scrutinized. Any negative side effects, such as added supplementary tasks resulting from the robot's introduction, must be accounted for.
Ultimately, the client must be able to estimate cost savings and confirm the Return on Investment (ROI). While some strategic deployments may accept a lower ROI in the short term, the primary value for clients—especially in addressing Labor Shortage—is operational cost reduction. Providing this value conclusively is what enables the client's decision for Mass Deployment.
The final stage involves the large-scale rollout of robots across multiple sites owned by a single client. This dramatically increases system complexity and raises significant issues regarding maintenance and management. Quality and delivery problems can surface during the introduction of numerous robots, and the process of persuading site-specific stakeholders remains a major challenge to overcome.
It's easy to lose track of how many years have passed by the time a successful Mass Deployment is achieved. Yet, at this point, only a single client's requirements have been met. There is still a long road ahead to satisfy the diverse needs of multiple clients.
Riibotics set the goal of developing an Industrial-Grade robot product from the very beginning. While many robotics startups focused on spectacular technology demonstrations, Riibotics concentrated on commercially viable product development.
Less than a year and a half after its founding, Riibotics successfully deployed its robot through a commercial project with a client, and the robot is currently operating in the client's production line. The positive reception has already resulted in a Letter of Intent for expanded deployment, which is targeted for next year.