In his recent contribution to WIRED’s Innovation Insights, Blue Prism Chairman Jason Kingdon explores the idea of a robot soccer team competing in the 2050 World Cup along with the “human teams” we see today. Jason examines whether robot players’ compensation would be similar to the likes of today’s professional players like Cristiano Ronaldo or Gareth Bale, and how this scenario is already playing out in today’s back office.
To read more, click here or check out the full article below.
World Cup Robots and the Economics of Robot’s Pay
During a warm-up robot football game (soccer, if you must) back in the mid-90s, one robot team was judged the winner on the basis that it had kicked the ball — literally once. These early, somewhat surreal matches, in which random motion and prosthetic air shots were carried out in the presence of a football, were the foundations of what has become the Robot World Cup.
RoboCup, as it is known, has since come a long way; 2013 saw 2,500 contestants from more than 40 countries. Games involve multiple passes, enough to impress Graham Taylor, and there are even real goals. The contest has also expanded to include different leagues for rescue and specialist robots. A headline mission is to have a robot team that can beat the human world champions by 2050.
So far, there is no league for routine administrative tasks — the domain of software robots, the subject of these articles. These virtual robots drive software devices in the same way users do, and so compete with humans for routine clerical tasks.
Would a 2050 robot team compete with human world cup winners on economic grounds? The answer seems trivially “yes.” If Real Madrid is pushing towards a billion dollars for its squad, robots will come in virtually free in comparison. The point being that the R&D cost will be sunk; once a robot Gareth Bale or Cristiano Ronaldo is built (at whatever cost), mass production will take over.
A similar issue, albeit a less glamorous setting, is playing out in the clerical back office. Software robots come in at about one third the cost of an offshore employee. The question is how does this compare, economically, to a person doing the job? Well, firstly, how much work can the robot do? Here, the comparison that is used most commonly is a 2×1 productivity gain. That is, a robot does the job of two people. This begs two questions: why so low? And why so high?
Why “so low” is a function of the robot mimicking a human. It carries out a task, after training, like a human. So unlike traditional IT, it is not attempting system integration — it simply automates a process flow using existing infrastructure (as is). This is why it is so light and fast to deploy. In many instances the robot will also be speed-restricted to only go as fast as existing infrastructure allows. Think of a robotic porter in a busy airport — it minds the traffic.
The big difference is that the robot can work all day, with no breaks, and only stops when required to — for example, at a nightly batch. All in all, this gives a simplifying assumption of a 2×1 lift.
Simple strategies can improve this. The robot can switch tasks to avoid downtimes and to keep workload up. Simple flow optimization can get us 4x or 5x productivity gains.
Of course, the robot could also drive much faster if the infrastructure allows. If work is re-organized to increase flow, there is no theoretical limit; certainly 10x or more can be achieved. Which, in terms of robot pay (i.e., the human loaded cost equivalence), means the robot is delivering 1/6 to 1/30 of the human cost! Notice, this still does not take into account human-economics such as – hiring, locating, housing, taxes, security, motivating, firing, error correcting, etc.
2×1 is therefore pessimistic. Why not higher? This is harder to appreciate and relates to the underlying nature of work as currently organized. To prevent humans going insane, work is often organically formed: tasks are broken up for interests and tolerance; switching becomes a defense and is tolerated so long as throughput is maintained — even a sprinkling of judgment is added to give spice.
Peter Drucker suggests the 50x productivity improvement in manufacturing over the 20th century was only achieved through hard work and creativity in reorganizing tasks to take full advantage of automation. This process is just starting in clerical terms.
Robots effectively empower operations teams to be creative and start to think of work as finding opportunities for automation, rather than just carrying out tasks. The results are that some organizations look like the early robot football squads, whereas others act more like they are celebrating the 2050 robot victory.
Factors underpinning performance relate more to the creativity and drive of senior teams than the players, or machines. In other words, the humans are decisive.
The best organizations seem to change working culture to embrace results. The teams act as leaders of robots and see automation as the goal. If you can train a robot to complete a task, you have achieved a durable prize — certainly a gift that keeps giving. Like the 2050 robot squad, the costs are sunk, and the standards and quality are fixed (robots don’t even age).
The other side of the 2×1 productivity gain is of course the margin gain. If costs have shrunk to 1/6 of a person, or most likely much better, the savings have become margin. Even assuming heavy robot overheads, the fact is that in most people businesses the costs are larger than returns, which implies a massive 100 percent or more lift in margins. This is the trophy. The clerical World Cup. For team intensive sectors, such as the Business Process Outsourcers running back offices, it is the chance to move up a division (or at least columns) in the Wall Street Journal; transforming people organizations into, effectively, technology multiples with up-skilled staff mixed with robots. The leadership could even achieve payoffs that rival Cristiano Ronaldo from very grateful shareholders.