Robots at Work: Domo Arigato, Partner Roboto
Industrial robots have long been manufacturing’s simple-minded caged beasts, safely fenced off from their human counterparts, toiling away at the most basic, dull, dangerous, or dirty work of production.
Today physical robots are being sprung from their cages. But this isn’t the stuff of dystopian science fiction. Thanks to significant advances in software and infrastructure, these increasingly capable and intelligent machines are safely working side by side – and often in close collaboration – with the flesh-and-blood enterprise workforce. Machine vision, natural language understanding, and other artificial intelligence (AI) capabilities are enabling robots to see, hear, and make sense of the world around them.
These more autonomous robots are fit for collaborating and handling one-off tasks, not just simple programmed repetition. “They can now safely occupy the same spaces as humans, and that’s why we’ve begun seeing so many new applications for robots in recent years,” says futurist and consultant Steve Brown. “AI enables robots to navigate the physical world and make more intelligent decisions about how to do so.”
Annual spending on robotic hardware, services, and software is expected to be US$128.7 billion by the end of 2022, according to IDC. This growing investment will drive increased innovation, lower costs, and greater adoption.
Meanwhile, a vibrant open-source community has emerged to collaborate on robotics software – including platforms such as the Robot Operating System – and hardware projects that are helping both researchers and companies develop prototypes and build robots better, faster, and more economically.
Meanwhile, the consumerization of AI systems and physical robots – from the Alexas and Siris of the world to social robots for healthcare – are making the sci-fi concept of robotic companionship seem normal.
But perhaps the biggest breakthrough is that robots no longer need to act like automated turtles, always carrying everything they might need everywhere they go, thanks to cloud computing and the continuing development of Internet of Things (IoT) technologies.
“When a robot encounters a challenge in one location and figures out how to overcome it, it can tell all its buddies almost instantly, and they learn together,” says Brown. These newer machines are more easily programmed and available at a fraction of the cost of their predecessors, making them accessible for even small to midsize businesses.
“The advent of a lot of powerful technologies that are increasingly cheap and ubiquitous means that we’re developing lots and lots of different types of robots that do lots and lots of different things – whether it’s a robot in a corporate setting or the Roomba that’s running around your house,” says Matthew Griffin, futurist and founder of consultancy 311 Institute.
New roles for robots
Capable of dealing with increasing variability and complexity of tasks with greater efficiency, these robots are being designed, developed, and deployed everywhere. Crawling robots and flying drones are inspecting otherwise difficult- or impossible-to-access pipelines and power lines.
Wheeled, collaborative robots are shuttling goods through warehouses to human shippers or being plopped on sidewalks by human drivers so that they can complete the last mile of delivery. Human cleaning crews are deploying giant robotic arms to swab ships and wash skyscraper windows. Super-brainiac robots using advanced analytics are helping surgeons in the operating room and enabling more precise and sustainable approaches to farming.
It won’t be long before gangs of intelligent and connected machines work in concert with existing enterprise systems and human workers to transform the way things get done throughout the enterprise, says Teresa Tung, managing director of Accenture Labs. Improvements in mechanical, electrical, and software components will enable these machines to operate in more semi-structured and unstructured environments. Soon, organizations will be able to program and integrate robots much in the way they do any other IT system, says Tung.
As the field expands, “The first movers will have a big advantage,” Tung says. Now is the time for companies to explore the potential value of robots in their organizations and develop their strategies for moving from experimentation to fully integrated robotic capabilities. In response to the COVID-19 crisis, for example, Accenture has been exploring how collaborative robots that leverage machine vision, deep learning, and automation can perform and sort environmental tests – such as inspecting petri dishes – to increase accuracy and speed, allowing skilled humans to focus on higher-value tasks. “This need for automation through robotics is more critical now than ever to augment a human workforce,” Tung says. A wide variety of new applications of robots is emerging across multiple functional areas and industries, offering a glimpse of the possibilities.
Robots in the enterprise: Where to begin
The best way to begin figuring out where to use robots in the organization is to map out the four or five most important business processes and break them down into detailed tasks, says Brown.
Then companies can start to ask the right questions: Which tasks are best performed by a human? Which are best accomplished by a physical robot? Which are best executed by nonphysical, AI-enabled automation (for example, algorithms)? Finally, which tasks – and there will be many – will be best served through human–robot collaboration?
Humans may still excel at creativity, critical thinking, empathy, or entrepreneurialism. Physical robots’ current assets include endurance, speed, and repetition, and they can be useful where superhuman strength or durability is required.
After answering those questions, business leaders will have a better idea of the makeup of their human–robot teams of the future. Then the real work begins: integrating those teams.
Companies will need to develop an integrated strategy that considers some major changes, says Tung.
As enterprises have become more digital, the volume of ongoing (and often tedious) IT management has increased exponentially. Much of that work is a perfect fit for physical robots.
For example, robots can serve as turbocharged gofers for data center repair experts. They can hunt down parts within massive data centers (where round trips can take humans 30 minutes or more), making the trips somewhat faster than employees can and freeing workers up to focus on higher-level tasks. They can also perform the physical testing of servers and other equipment. “We imagine a future where data centers are staffed almost entirely by robots, to the point where a human entering that environment might actually slow things down,” Tung says.
Robots can also take on tasks that are so tedious and time consuming for humans that companies have been forced to take shortcuts. Utility companies, for example, have so much equipment to test that they can only afford to examine a sample. Robots can perform all the tests – tens of thousands of them, often involving little more than pushing buttons – in a fraction of the time humans can and at a fraction of the cost, so companies can make certain that all the devices are working before deploying them.
Robots are also taking over tasks that require greater intelligence and fine motor skills – like picking goods. Germany’s Magazino is producing perception-controlled, mobile robots that use 2D and 3D cameras to identify objects, securely grasp them, and transport them to their destination. This results in extended operating times, a smoothing of order peaks since companies can more easily scale up and down the robot force than the human workforce, and a reduction of operating and process costs.
Robots can also become specialists in certain tasks at lower costs. Magazino’s TORU robot is made specifically for handling shoeboxes and helps to cost-effectively address some of the unique challenges of the footwear industry, like the decreasing of order sizes, Monday peaks, and high return rates.
At the other end of the supply chain, companies are developing prototypes of robots that can deliver everything from packages to pizzas. In fact, Lux Research predicts that automated last-mile deliveries will generate up to $48.4 billion in revenue and account for 20% of parcel deliveries by 2030.
That will require robots to navigate what can be a very complex and disordered world. In February 2020, Mountain View, California-based Nuro became the first company to be granted approval for a self-driving vehicle exemption and has partnered with grocery stores to provide unmanned deliveries in Houston and Scottsdale, Arizona.
Want your pizza delivered to a fourth-floor walkup? No problem. Unsupervised.AI’s four-legged delivery robot, Aida, uses deep reinforcement learning techniques to navigate staircases, street curbs, and other obstacles. Co-founded by a former Lyft data scientist, the company plans to build a network of drivers (as the ride-sharing service did) to ferry the robots to street corners to embark on their delivery adventures.
Surgery has proven to be the must-have app for physical robots in healthcare – one that is already improving outcomes and saving lives. Intuitive Surgical, which grew out of a military program in the 1980s for conducting remote surgery on battlefields, has more robotic surgery systems installed than any other, with 5,582 systems in place by the end of 2019.
The company’s cyber-sawbones help doctors perform abdominal, colon, gynecological, heart, and neck and head operations. Its latest innovation is a robotic instrument that uses CT scans to create a road map for lung procedures and then moves through the lungs’ airways with a tiny, flexible, snake-like catheter to take biopsies of lesions that are otherwise difficult and dangerous to access.
As consumers have grown accustomed to interacting with disembodied, AI-driven applications from call centers and chat bots, as well as Alexa and Siri, it’s only natural that physical robots would be next. Pepper, the four-foot tall humanoid robot from SoftBank Robotics, which has invested heavily in the development and marketing of the platform as a customer-facing android, has quickly become the go-to social robot for customer-facing applications and experimentation.
HSBC employs Pepper at some of its retail banking locations, including its flagship Fifth Avenue branch in New York City. There the robot works to attract customers by engaging with pedestrians near the entrances, answering basic questions, and directing people to the right human for further help. In December, the company opened its first café where robots assist humans in serving customers in Tokyo. During the COVID-19 crisis, a supermarket outside Cologne, Germany, deployed Pepper to urge customers to keep their distance from one another and not buy in bulk.
Knightscope’s K5 security bots, which have been described as a cross between Star Wars’ R2-D2 and Doctor Who’s Daleks, are built to keep public areas safe by rolling through open areas, halls, and corridors looking for and reporting suspicious activity to human supervisors.
These five-foot, 300-pound mall cops upload what they observe to a back-end security network using 360-degree, high-definition, low-light infrared cameras. An audio event detection system can also pick up on activities like breaking glass and send an alert to the system.
These bots are not shy, either. They have built-in microphones and speakers that can be used to communicate with passersby.
The K5s are popping up everywhere, from shopping gallerias to office buildings. The under-development K7 is an all-terrain version designed for clients with large outdoor facilities, such as airports, utilities, and gas distribution centers.
Ripe for robots
The global agriculture industry is contending with significant labor shortages, as fewer people are willing or able to do the dirty and often dangerous work in the fields. After decades of steady decline, world hunger has slowly been on the rise since 2015, according to the Food and Agricultural Organization of the United Nations. Meanwhile, crop yields are already being affected by climate change and productivity must increase to accommodate the 800 million people that will join the global population in the next 10 years. These problems are ripe for robotic automation, and several innovations are already taking shape.
The precision required for fruit picking was once something only humans could deliver. However, California-based Abundant Robotics has built an automated apple picker that uses a vacuum system to suck the fruit off the trees. While it sounds like a low-level task, the robot picker actually must identify the apple, determine if it’s ripe, and then pick the fruit without damaging it.
In the United Kingdom, the Hands Free Farm project used automated machines to grow the world’s first arable farm remotely without any human operators in the driving seats or living, breathing agronomists on the ground. “Their challenge was to sow, tend, and harvest an entire field of barley without ever touching it with human hands,” says Brown. “And they did it.”
The team has since put a cover crop into the hectare to protect the land as it becomes a test space for further developments, including working on the tractor it used, integrating technology from self-driving cars, so that it can interact with its surroundings and drive itself.
Swiss company ecoRobotix has produced a solar-powered smart weeder that uses AI, cameras, and two robot arms to spray doses of herbicide on weeds but not crops, using 90% less herbicide than traditional methods.
Leaders of the pack
In the short term, early adopters of physical robots in the enterprise will gain an advantage by cutting costs, eliminating manual workflows and inefficiencies, and better mitigating risks. Looking further out, however, robot-driven automation will become a competitive requirement across many industries.
As robots continue to grow smarter and more collaborative, companies will need to figure out where they fit in the enterprise, not just from a technological or operational infrastructure point of view. They will need to consider how human and machine can work together from a human resources and management perspective as well.
If there is danger ahead, it will not be a result of the advancing robotic beings that are evolving and entering the enterprise but of not preparing adequately for this eventuality now.