Robots - it’s about the service, not the hardware

Robots that can rush in where people fear to tread are a major growth opportunity for innovators in robotics, AI, and autonomous systems - as explored in my previous report on nuclear decommissioning, space tech, offshore energy, deep mining, and UK developments in these fields.

But not every sector is a hazardous environment for human beings - at least, that was the case pre-Covid. The services sector and others that keep the customer clothed, fed, and supplied with goods also represent opportunities for robotics growth and innovation.

While the pandemic may have helped to soften the public's attitude to robots, by revealing that automated services could be a boon in industries such as retail, manufacturing, farming, energy, transport, and logistics, will people remain more accepting once things get back to normal?

Arguably, the key to long-term success in robotics is simply providing useful services, and not a focus on the hardware. The media's dystopian obsession with robots themselves, as though a Terminator will be sitting at your desk and stealing your job, has long been misleading. If nothing else, sci-fi economics rarely stack up.

When I was in the US in 2019, compiling a report for Innovate UK on robotics in extreme environments (in the pre-Brexit Expert Missions programme), the CEO of one US service robotics company told me that "no one makes money from selling robots".

He explained that most robotic hardware is really a loss leader for selling services into clients that need to automate repetitive, but critical functions - in his company's case, cleaning and maintenance. In other words, they are a means to establish new revenue streams for services companies, rather than a one-off hardware sale of an expensive, depreciating asset.

Factory production lines in automotive, aerospace, and electronics aside, there is some truth in that view. Robots that can run, jump, pick up objects, or do somersaults may keep people entertained on TikTok, but unless they can do something useful - and do it better, faster, or more safely than a human - then they are simply pushing the engineering envelope. That's important, but it's not enough when establishing a new market.

Not meeting expectations

At some point in the future, science fiction may move closer to reality: when autonomous machines mine asteroids, for example, or act as co-workers on Mars. Autonomous vehicles may eventually be adopted at scale in our cities, alongside pilotless air taxis and maintenance drones. But until then, most robotic applications will remain more down to earth. That means them doing things that people actually want.

This point has been proved in recent years by the shops, banks, transport networks, hotels, and restaurants that have grabbed column inches by putting robots in front of their customers, only to withdraw them later when human staffing costs actually increased, ventures fizzled out, or the public was baffled by machines that had less intelligence than their smartphones.

Unless they provided useful, profitable services, such robots were little more than marketing tools: machines that looked good in news photos but struggled to answer simple questions, or left customers jabbing at touchscreens. If nothing else, a century of science-fiction has raised expectations that few robots can yet meet.

Robots of every kind must provide a useful service - and keep providing it reliably. A boring autonomous cleaning vehicle or warehouse robot that works 24x7, slashes costs, speeds up business, and improves the environment for humans is a much better commercial bet than a flashy humanoid machine that leaves users frustrated at its costly, disappointing limitations.

The humanoid robots beloved of dystopian fantasists do exist, of course, but many are research platforms for future co-worker or care applications, while others might be impressive pieces of engineering, but struggle to find useful apps and practical applications to justify the expense.

Real world use

In the real world of profitable robotics, it's all about helping other industries keep their services running and their customers happy. Brian Palmer is CEO of Tharsus, one of the UK's leading advanced machines and robotics firms that meet practical, real-world needs.

Based in Blyth in the North East of England, Tharsus specialises in designing and building robotic systems that tackle real automation challenges, taking them from proof of concept to adoption and deployment in commercial environments. That's a significant challenge. 

According to figures from the company, Tharsus Group achieved strong growth in pandemic-hit 2020, increasing revenue by 40 percent year on year from £58 million to £81 million, while pre-tax profits almost doubled. So, it must be doing something right, though the figures arguably remain modest compared with the revenues of its customer base.

Clients range from the likes of DHL and Ocado to Rolls-Royce (aerospace) and BMW, plus smaller ventures, such as agricultural robotics specialist Small Robot Co, and CMR Surgical, the Cambridge-based unicorn that is all about putting life-saving robotics in hospitals.

Its OEDM (original equipment design and manufacture) projects for DHL and Ocado, for example, are about warehousing fulfilment robots - in Ocado's case, swarms of machines working on a grid to pick groceries faster than humans can. In this way, they help get food to customers quickly. Tharsus helped the grocery giant to design, scale, and manufacture the machines, and the client retains the IP.

Tharsus is helping BMW develop service robots within its own factories and warehouses - its expertise is in cars, after all. While collaborative robots (‘cobots') are often talked about, replacing dumb, single-task machines in safety cages with more adaptable ones that can operate safely among humans is a tall order. Especially when those dumb machines have helped the automotive industry thrive for decades.

Meanwhile, the likes of Rolls Royce have been hothousing robots for on-wing jet engine maintenance, including remotely. The vision is engineers in Derby servicing a plane on the tarmac in Dubai.

Why? Because Rolls Royce is itself now a services business more than a hardware seller, as so many hardware-centric companies now are. It is paid when planes are in the sky, powered by its engines, and not when they are on the ground being serviced.

Robots controlled by highly trained aerospace engineers minimise expensive delays - for airlines and their passengers, as much as for Rolls Royce - and keep the services revenue rolling in. So again, it's not about selling an expensive engine; it's about revenues from a reliable service. That's the new economy.

Strategic machines

However, Palmer doesn't accept that US CEO's assertion that no one makes money from selling robots - that's a "harsh view", he says, and clearly Tharsus is in the business of supporting its clients with technology development and manufacturing. However, he does acknowledge the underlying point that it's all about providing a useful, end-to-end service:

We talk about ‘strategic machines'. It's a bit of a wrapper, because if you look at what Ocado has done, and the robots we have built for them, it's part of helping them to create a service offering.

They've created a value chain that goes from the phone app or the Web browser, through to the PDA that a driver uses to see an order through to delivery. The robot is a small element of that big value chain, but also an enabler of that value chain. We provide one small component of the value chain in a robot.

It's not just about building a robot that can operate by itself, he adds; it's about spinning an ecosystem around it.

We're generating a labour and space saving, so that's a really simple ROI. But we're also generating a data flow that will create greater accuracy of supply and higher frequency of supply. So, there's a lot of secondary benefits, which all add up.

A hardware platform alone would probably give a marginal return on investment, but if you can improve the service offering of your customer, then suddenly you've completely changed the value equation, and that's when it becomes really exciting.

How does he see the future shaping up for companies in this space?

We're seeing more and more emergent companies with new business models that are aware of the automation opportunity. It's not about replacing jobs or saving a few percentage points on costs. It's about enabling new services and service propositions to be developed, whether that's for the consumer or a commercial customer. That's where we see the most exciting developments.

Is interoperability the key to widespread adoption?

Some in the industry believe that interoperability among autonomous mobile robots (AMRs) will be the key to more robots being adopted in factories, warehouses, and distribution centres. Until recently, robot fleets from different vendors lacked a common means of coordinating activities, sharing data, or adopting the same operating conventions.

Last month, Boston, Massachusetts based non-profit MassRobotics announced the release of what it claimed is the world's first open-source interoperability standards for AMRs: version 1.0 of the MassRobotics Interoperability Standard. This will be road tested at a FedEx warehouse in Tennessee this year, where robots from Waypoint Robotics, Vecna Robotics, and others, will be collaborating in the same location.

Functional and practical standards are a critical next step for robotic automation," said Tom Ryden, Executive Director of MassRobotics, at the launch.

According to figures from supply-chain market intelligence company LogisticsIQ, the global AMR and Automated Guided Vehicle (AGV) market will hit revenues of $14 billion over the next five years, with an estimated 270 vendors or more focused on the manufacturing and logistics spaces alone. AMR adoption is forecast to see a compound annual growth rate (CAGR) of 45 percent in the same timescale.

However, divide that $14 billion among 270 companies and you end up with revenues roughly the size of Tharsus, so these aren't businesses that are likely to be unicorns themselves. That mitigates against the robotics market ever being about selling expensive, mass-market service robots to the public, for example - particularly as it is far from clear what they would actually be for.

The likes of Amazon might take a loss leader approach - as it did with its Alexa-powered devices - but there would have to be an obvious way of using robots to sell Amazon goods, content, and services. And that means finding a discrete use for them that can't be met by a tablet, smart speaker, TV, or doorbell.

Back in the realm of service robotics for industry, Palmer suggests that interoperability among competing manufacturers isn't the big issue in growing the markets that Tharsus operates in.

This is where we're getting into the world of robots that are designed around a very specific service offering, versus industrial robots, say. If you take your car company, your paint shop, or your body shop where you are welding car bodies together, you've probably got a range of manufacturers and some fairly common platforms.

Whereas if you create a new platform, yes there is integration with the outside world and maybe different technology vendors in there. But we're designing application-specific devices, so you wouldn't be taking something off the shelf from someone else to do the same job.

We're creating new devices for our clients. Yes, we've got some common building blocks, drives, and sensors, and so on, but the configuration is unique to the customer. When we create, we are generally trying to create a new service. You don't necessarily mix fleets.

Looking ahead

So, what of the future for Tharsus, and companies like it? The pandemic has been the proverbial wake-up call for a range of different industries, says Palmer.

A lot of companies that were moving into this space thought they had time on their side but are suddenly realising how quickly the world can change.

Climate change and the need for sustainable, green solutions are part of the challenge, as well as looking at how industries such as agriculture and food production, for example, can be made more resilient to acts of God or the scant availability of migrant workers.

But it's not as simple as just building a robot to plough and irrigate fields or pick fruit - though these are all areas of development in agritech, alongside the use of drones and sensors to monitor crops and livestock.

The amount of processing power that such a robot would require to do the same things as a human is extreme. And there are many subtleties in the way a human picker will work.

I think the real opportunity is more in optimising the crops themselves - how do you weed, how do you feed, how do you sow? All those things improve the overall outcome, rather than just carry out a simple task. It's about improving the efficiency of farming at a holistic level.

My take

Once again, it's about providing a service that understands both human and industrial needs. It's rarely about saying, "Here's an expensive robot. Set it to work and fire all your staff."