Robots as a service: The Pepper example

Softbank’s Pepper brings the march of the robots closer to reality with a big step forward. But it also highlights the long road ahead.

The popular robot – Softbank sold than 7000 units in 7 months in Japan during 2015 – is arguably the world’s leading example of the consumerisation of robotics in action today and should be on any business and IT leader’s list of emerging tech to evaluate. Even if Pepper isn’t the most advanced in terms of technology, and it isn’t by a long stretch, it is supported by the most progressive, ecosystem-based business model. And this latter factor is probably more important to Pepper’s future success.

Pepper has succeeded in moving robotics out of the industrial and manufacturing spaces where they have performed admirably in driving greater automation and precision, into a world of human interaction both at play and work. A space where many – including many venture capitalists and well-resourced enterprises – feel robotics and artificial intelligence have the potential to replace many jobs or tasks, but which have struggled to take hold to date.

But while the approach Softbank has taken with Pepper is to be applauded and introduces an exciting new dynamic to the market, this new dawn in robotics where humans are replaced at a grand scale is some way from breaking if Pepper’s drawbacks are any indication.

Some of the positives we see in Pepper include:

  • “As a service” use of robots: Pepper is available to organisations in Japan (and only Japan at present, although overseas expansion is intended in future) via a 3-year lease contract with off-the-shelf applications and services. Softbank or its partners take ownership of break/fix services. This approach should allow more organisations to try Pepper as they don’t require a significant capital or skills investment up front.

  • Pepper is well known: Very well known as a result of a dedicated marketing campaign. This means it has much needed traction with those not readily familiar with the world of robotics.

  • There is a marketplace that forms the basis of an ecosystem: Softbank recently announced a marketplace called RobotApp Market for Biz which includes off-the-shelf applications and services. At a recent event we saw applications spanning: healthcare, education, retail, and tourism. There is also a software development kit (SDK) for developers. This will help Pepper become the first robot that has a true ecosystem supporting it.

  • Early adopter reference cases are abundant: We have seen Pepper in action with Tsutaya Electric, Nissan, Nescafe, Yamada Denki, Nestle, and Loft to name but a few.

  • Pepper’s performance is generally acceptable: Although Pepper occasionally struggled to provide a good experience due to the throng of people at ‘Pepper World’ in Tokyo recently, its performance is for the most part acceptable.

While the above positives are welcome developments in this emerging era of robotics (and AI) some of the downsides of Pepper today include:

  • The novelty wears off pretty quick: If you have ever interacted with Pepper the experience is pretty good. The performance of the robot is relatively seamless and it can provide a reasonable experience. But for pretty much every application the robot part of Pepper – i.e. everything except for the screen located on its chest – is effectively an accessory. For education, healthcare, retail, and tourism applications all of the experiences can be just as easily provided via any screen type sans the robot. Considering Pepper is available on a three-year contract you have to ask whether it’ll still be offering benefits at the end of the engagement. Or, will Softbank provide hardware upgrades throughout the contract?

  • The total cost of ownership (TCO) of Pepper is unknown: Although Softbank or its partners takes care of hardware issues, there isn’t enough information on what the breakage or failure rate is. Moreover, the labour and other costs (such as electricity) associated with implementing Pepper at a given site (store, school, hospital, etc) are not readily available.

  • Pepper’s early applications are singular in purpose: In all the use cases presented and in our interactions with Pepper, it appears that the robot is only used for one application at a time with little effort to allow it to switch easily to other modes or take action autonomously. For example, it would seem appropriate to allow Pepper to start being useful with basic tasks when not interacting with a customer – such as cleaning the floor or measuring temperature or undertaking security/safety operations. However, at present our experience with Pepper is that as soon as there are no people to interact with, it goes into a sleep mode that looks as though it has nodded off.

  • Pepper needs human support: In many situations Pepper still requires a human to support the interactions it offers. Whether it is to provide the actual service being offered, or to answer any questions customers may have during or after an interaction, organisations that adopt Pepper will need to plan and budget for the labour required to support its deployment.

  • It is not made clear up front how Pepper handles data: While we assume Softbank and the application developers will be able to answer this question, it is a concern that data handling is not dealt with up front in a transparent manner.

  • The first stages of innovation are limited to Japan: The decision to only offer Pepper domestically at first is a typical approach for Japanese developers of technology and in many ways makes total sense. But it is also an approach that continues to result in what is known locally as the “Galapagos effect” – i.e. technology that works well in Japan but isn’t successful once taken overseas. Softbank is risking its first to market position by not immediately offering Pepper in markets outside Japan and also is missing feedback from non-Japanese users in the all-important early adopter stage of the market, which will be critical to ensuring adoption outside of the home market.

  • The market place is just for Pepper: While the market place is a fundamentally smart approach, it could be improved by opening the platform up to applications for other robots. Thus allowing organisations to deploy multiple types of robots matched with any type of application they desire, and all as a service.

While it is easy to find fault with Pepper today and it is clear that there is still a long journey ahead before robots like this replace human interactions, the advances that Softbank have made with the business model and consumerisation of robotics for times that involve human interaction should not be underestimated. It is a significant step forward that has popularised or at least proven the possibilities of robots at work outside of industrial use cases. We encourage all organisations to investigate the potential of Pepper-style deployments as part of their emerging technology discovery and innovation programs.