Mr. Robot comes to town

A pair of robotic arms reaches for a row of canned drinks on a store shelf. Closing one steely palm around a can, the robot firmly picks it up and places it on another rack. It then repeats the action over and over, growing more precise each time.

Just behind it, a team of people orchestrates every move through connected virtual reality (VR) gear. Across the room, several groups of humans huddle around their own robots, practising the same controlled motions with various grocery store items.

“With each movement, the robots are continually learning through data from their sensors, cameras and video feeds,” Edwin Yap, founder of Mr. Robot, tells StarBiz 7 on a tour of the company’s recently opened data collection centre.

“It’s sort of like building muscle memory into them – they’re being trained to know how their robotic joints should react in different circumstances.”

The data collection centre, which simulates real-world retail environments to allow robot learning using artificial intelligence (AI), was set up with an investment of around RM10mil, according to Yap.

He says building the facility was necessary to address the substantial data gap within physical AI – a field that involves AI systems embedded in machines to enable them to perceive and interact with the physical world.

“One of the biggest bottlenecks for physical AI is the scarcity of data. Compared to large language models like ChatGPT, which have extensive resources to draw from, there is very little available data that can be used to train machines in the physical functions required to help solve today’s business pain points.”

Mr. Robot was first formed as a joint venture between Jemco Group, Yap’s family business specialising in shelving and racking solutions for retail and warehouses, and MR DIY Group.

Jemco’s years of serving retailers provided an in-depth understanding of the industry’s longstanding challenges, such as labour shortages, service inconsistency and limited scalability.

So when AI began its meteoric rise, the team saw an opportunity to harness the powerful emerging technology towards solving the industry’s problems.

Mr. Robot today deploys AI-powered robots in commercial environments such as

retail stores and manufacturing facilities. Within the retail sector, Yap says the company has focused on three areas of application.

This includes brand events and customer service, where humanoid robots are able to engage with people through public interactions, performances and supplying in-store guidance.

Its models have been deployed to this end in retail outlets, including a MR DIY store opening event that resulted in a 20% sales lift and less-than-a-year return on investment, he shares.

Beyond this, the business is developing its own proprietary software to enable its robot fleet to carry out inspection and patrolling tasks.

On the more technically complex end of the scale, robots can also be utilised for stock replenishment, alleviating one of the most repetitive and labour-intensive aspects of retail operations.

At the same time, the company is seeing a large influx of interest from the manufacturing sector, Yap says.

“It’s a big gamechanger for manufacturers. In the past, robots in industrial settings were programmed to only do a single task repetitively, but a humanoid robot can carry out many different functions with the same level of investment,” he adds.

In addition, robots are able to work 24 hours, increasing efficiency while taking over repetitive and potentially injury-causing tasks such as continuous heavy lifting.

Enhanced speed is also coupled with improved accuracy and consistency. Yap shares an instance where its AI-driven robots achieved a 99.98% task success rate when deployed in a manufacturing environment across seven consecutive days, as a result of reinforcement learning combined with extensive study of existing operational workflows.

While Mr. Robot envisions the robots eventually taking a range of formats based on different requirements, it sees humanoid robots as the most effective way to introduce robotics into existing work settings.

“At the moment, almost everything we use to accomplish tasks, from tables and power tools to entire factory setups, are centred around human ergonomics. For this reason, designing humanoid formats allows the integration of robots without requiring too much change to an environment,” he says.

Currently, the company’s business model is largely based on outright purchase and hardware sales, alongside the provision of engineering support services for clients. However, Yap sees this evolving into a licensing or robotics-as-a-service model, with the business guaranteeing robot uptime and value-add to a facility.

“We’re doing extensive research and development (R&D) activities aimed at creating packaged solutions that are more scalable. Eventually, we plan to go from being asset-heavy to asset-light, which will enable us to improve our margins,” he says.

Notably, the founder considers Mr. Robot to be less driven by robotics and more by the AI technology behind it.

“Our eventual goal is to have more hardware-agnostic solutions, which can be used with different types of robots. The most important objective for us is helping our customers to achieve their business outcomes,” he says.

For now, the company is prioritising the expansion of its engineering resources and workforce to boost its R&D capacity. Additionally, much effort is being put into widening local awareness and understanding of ongoing developments in AI robotics.

Part of this awareness push is the upcoming launch of its AI-focused robotics store in Mid Valley City, Kuala Lumpur this November, which will introduce high-end AI lifestyle gadgets from abroad into the Malaysian market.

“AI robotics is still a relatively new field, especially in Malaysia, so we are seeking to bridge that gap,” Yap says.

“We’re also working very closely with policymakers such as Sirim and local universities, engaging with educators and participating in guest lectures to help bring up the ecosystem in the country.”