Roboman Satoshi Amagai

From Walkman to Aibo to ... who knows? A look at what's next from the president of Sony's Robot Entertainment Company.

Media interest in Aibo -- Sony's quadrupedal, anthropomorphic robopooch -- was little short of hysterical after the product hit the streets in a blaze of publicity and light-speed sales in June 1999. The statistics of the initial limited offering are already the stuff of legend -- 2,000 shiny $2,500 Aibos sold in the US in the first four days, while the domestic novelty seekers snapped up the initial run of 3,000 in less than 20 minutes. Since those early days, Sony has spun off its robot division into an autonomous entity -- the Entertainment Robot Company -- under the stewardship of Satoshi Amagai, and it's established Aibo as a permanent fixture in its product line. Indeed, Amagai has been widely quoted as saying that he hopes to see Aibo become one of the main pillars of Sony's business. As of April 2001, more than 95,000 Aibos have found new homes around the planet.

While undeniably a success story, Aibo is not without its critics. Noted robotics scientist Takeo Kanade, the director of Carnegie Mellon University's Robotics Institute, offered his view. "Aibo's 'cuteness' comes from its inability to do certain things, rather than its ability," he says. "The machine is not complex enough yet to show that [kind of] complexity in the pattern of failures, let alone to do the right things."

Nonetheless, Kanade sees no shortage of economic potential: "I think this entertainment robot market will be real, in the sense that they're dealing with 'nobody-knows-the-right-answer' or 'nobody-cares-about-the-right-answer' problems." That's an academic's viewpoint, but what of the man driving Sony's new business division? Amagai has something of a track record at Sony, and is often cited as the driving force behind the success of the original Walkman -- something he was at pains to clarify for us. His background in AI at the Tokyo Institute of Technology led him into the TV and video group of Sony Corporation and from there to defining audiences and forecasting demand for Walkman and new media teletext services, respectively.

Various spells in the US saw Amagai create Sony's four-zone (Japan, Asia, US, Europe) business system and apply it to the global television market, as well as establish a manufacturing infrastructure in San Diego. Most recently, he has been GM of the Home Network Company, charged with integrating Sony's diverse business strategies.

Amagai met with journalist J. Mark Lytle at the year-old Entertainment Robot Company's Tokyo headquarters to share his understanding of the current market for consumer and entertainment robots -- and his vision for the road ahead.

When you were a student at Tokyo Institute of Technology, did you ever imagine you would be masterminding an army of electronic animals?
I was always very interested in making robots, because, as you know, in Japan there's a history of friendly robots, such as Astro Boy; however, after entering TIT, I realized my limitations and focused on AI, rather than on mechatronics. I was also strongly influenced by the intelligent computer HAL in Stanley Kubrick's 2001, a Space Odyssey -- in fact, I've seen the movie five times.

You were in large part responsible for the success of the original Walkman. How do you compare that challenge to the task of making Sony's robots as successful?
That's not necessarily correct -- my original TV/video group often covered other product areas, including the Walkman. I was deeply involved in forecasting the potential target customers for the Walkman, which was a completely new category at that time, but it was just one of the products I was involved with. Because the Walkman was such a hit, people frequently mention it when they refer to my role. I actually did more forecasting for TV and related products.

My challenge now, when compared with the Walkman, is very different. Even though Walkman represented a new category, it still was an extension of existing products. In other words, it was still for listening to music. Aibo had no benchmark and involved creating a new market from scratch -- that's the difference: expanding an existing market, or creating one from zero. It's my greatest challenge.

Of all the products you have been directly involved with at Sony, which has been the most interesting? Which has been the greatest challenge and which the most satisfying?
I would say the current Aibo is most interesting -- interesting and challenging are two sides of the same coin. In the US in the early '80s I was concentrating on TV product localization, and in 1985 we introduced a flagship television, much like the Toyota Lexus is a flagship car. We created the XBR category, which stood for eXtraordinary, Black tube, and Remote. It was picked out by many magazines as a benchmark or reference point for excellence and is still referred to almost 20 years later.

How does Aibo fit in with Sony's other products? Where do you see its strongest synergies within the broader company?
Because Aibo is unique, even within Sony, we are independent. You can have too much synergy, of course. From the hardware viewpoint, Aibo is really Sony hardware in that it's creative, amazing, and fun to play with. Some of the key features and appealing points are shared with other product groups. Inside Aibo is a core PC element -- the email reading ability is the fruit of our cooperative work with the Vaio group. In that sense we are not isolated; rather, we try to share some of the same values.

What are the economics of Aibo?
It's pretty difficult to give you figures. Reason one is that we became a company just a year ago; before that, we were a special project, so total figures were not necessarily filed. Of course, the first trigger in Aibo's development was in '93, so are we supposed to include all the R&D activities from then in our balance sheet? Forgetting about the background, I'll say that current business is not so bad, but not so good either. Because it is a unique product with no head-to-head competition, Aibo lovers buy Aibo. Affluent people in particular don't care about the economy, but in my opinion, second adopters are waiting for a better economic situation. The world economy is very sluggish right now, so maybe the purchase of hobby items is not a priority. Our business situation is not better than we anticipated, but not that bad. We're not necessarily that rich!

Are sales disappointing?
They're within our forecast range. But when we meet with top management, they are very warm toward us and they always encourage us to do something new, or creative, or more fun, rather than telling us to stick to the bottom line.

What are the predictions for future growth?
We'd like to see Moore's Law apply to give us a doubling of business every 18 months. I don't think it's a dream, because the Japanese market is really almost a cult, and in the US and Europe there's still a very low awareness, so we need to increase that awareness abroad.

Aibo was launched using an interesting business model -- the limited edition. What bearing has this approach had on the success of the product line?
It was just test marketing -- we didn't really intend to sell it forever. The first two occasions were limited quantity marketing, then the third became a limited period marketing. Then we seriously started to think about forming a real business.

What are your plans for Aibo's evolution?
Even now, Aibo customers are very diverse -- some see it as a pet or an emotional friend, so they don't want to know about the computer inside. Other customers regard it as a very high-tech, interesting PC gadget. They like to take it apart, look inside, and even program it by themselves. One option is to expand the lineup to better satisfy those two different kinds of customers.

Color is also an option -- currently in Japan, the gold Aibo sells best, but in Europe it's silver. Even in color, each region has its own preferences, but we can also introduce different software or accessories. If the market grows, regional models are a possibility -- perhaps an American version or something like that.

Also, a "connected" or networked Aibo (a wireless LAN connection is already possible) that can be synchronized remotely with a software Aibo will allow more opportunities for entertainment -- Dokodemo Aibo (Aibo anywhere), perhaps.

Inside Aibo is a PC, so we have to find the most suitable applications for it. Some procedures are better done by PC; Aibo isn't going to be a word processor, but it may read your mail for you, perhaps even with a local accent.

As far as price goes, one alternative is to keep the price as it is and insert higher features and technology. The other is to keep the same features and lower the price. That's something of a choice between zero and one, but I think we can find a compromise. I can't say right now which direction we will take, but we are discussing our options at the moment.

Those are all possibilities -- are there any probabilities?
It depends on the size of the market, of course, so we have to consider economies of scale. Currently, the Japanese market is number one by far, accounting for over three-fourths of all Aibo sales. We have to prioritize and can't necessarily satisfy all markets immediately. That's what I'm facing.

How about third-party products for Aibo?
One application allows it to read books to you or your children, while another allows it to read out email. There are also other applications that let Aibo play cards with the user or with another Aibo.

Almost every day we get inquiries from outside developers, but we have to consider the business potential. So far almost 100,000 Aibos have been sold -- some people say this is a large figure, but, to compare it with the established base of, say, game consoles, it's still a small number. Unless you hit 1 million, outside parties don't get serious.

Apart from Aibo, what are the most interesting products in this sector?
From the viewpoint of hot sellers, Furby and Sega's Poo-Chi were big hits, even though they're really just toys. I guess they made more than $400 million worldwide. In a way, it's a good benchmark for us, but they're very different from our product. As of yet, there's nothing over $100 except Aibo. In fact, it's now in the Guinness Book of World Records as the fastest selling entertainment robot. My guess is many potential competitors are watching us, as was the case with the Walkman, which was out for two years before serious competitors emerged.

Did cheaper competitors dilute the brand value of Walkman?
I think the reverse -- they meant Sony was seen even more as the flagship brand. Although we did eventually market cheaper versions, especially in the US, the brand image of Walkman was never eroded.

Does the same thing apply to Aibo?
I think so -- now, every time the press or TV feature robots, it's always Aibo, isn't it? Frankly, we welcome competitors to grow or expand the market together.

The Walkman people may object to this comment, but compared to that, the hurdles to entering this market are higher. As I said, to make an attractive robot, three factors are necessary: very competitive mechatronics, AI, and entertainment sense. Maybe some competitors have one or two of those elements, but not all three, so it may require a joint venture or international consortium to enter the field.

What kind of robots are we likely to see in homes in the future?
I guess they may divide into three categories -- one, like Aibo, will continue to concentrate on entertainment. Second, task-oriented or domestic robots may evolve to help with domestic chores. Third, a medical robot for such things as home nursing is possible.

As of now, Sony has no intention of going into the second two categories -- we know we're good at entertainment, but we don't know if we're good at nursing or domestic chores.

Also, I see more intelligent robots -- the current Aibo uses voice recognition, but maybe the next challenge is not voice, but speech recognition or perhaps even a full dialogue. I don't know if it's attainable in the next 10 or 20 years, but emotion is one of the goals. At this moment I've no idea, but for me personally, I don't believe in the phenomenon of the character David in (Steven Spielberg's movie) A.I., where a robot wanted to become a real human.

Are there any potential dangers in marketing intelligent or emotional toys?
People's perception of robots is very different in Japan and other countries -- I'd say that outside Japan there's a tendency to regard robots as servants. When they break Asimov's Laws of Robotics (see box) and fight humans, that's the starting point for many movies and novels, so I don't think I can say there's no danger! I often talk with our godfather, Dr. Doi [Toshitada Doi, executive vice president of Sony], about the progress of AI taking time and being heuristic, so you never know how soon the great leap to the next stage will come.

On a lighter note, can a lump of plastic and silicon ever replace the beloved family dog?
Two comments -- we won't necessarily stick to animals; I don't know what will come next. Secondly, we didn't try to replace live animals, but Aibo has been popular with many Japanese customers who cannot have pets. It's the customer's choice -- we haven't been pushy in that area at all.

In a nutshell, what's Sony's outlook for entertainment robots?
Our near-future goal is "living with intelligent robots" -- we endeavor to make that happen. @