Consider this: a British technology company launched a new software category in the late Noughties. Its products are now used by 1,400 Fortune Global 5,000 organisations in 130 countries, and its market capitalisation (at the time of writing) is over $800 million – more than Google paid for DeepMind.
The name of that company ought to be on the public’s lips, but for some reason few people have heard of Blue Prism, prime mover of Robotic Process Automation (RPA). It’s a puzzle.
When you consider that Blue Prism was born in what UK politicians call ‘the Northern Powerhouse’ – Warrington, between Liverpool and Manchester – and in April 2019 the value of its shares was north of $2 billion, the mystery of its low profile deepens.
Indeed, Executive Chairman Dr Jason Kingdon claims that on the company’s £21.1 million ($27 million) IPO in March 2016, Blue Prism received no press at all, proving that it’s possible to float a company without the broadsheets’ backing. Not that the company wanted to be ignored.
Now based in London, Blue Prism is a multi-national with US presence in Austin, Miami, and Chicago, but its shares have clearly been on a rollercoaster ride in 2019. One dip coincided with its £80 million ($102 million) acquisition of Automation-as-a-Service provider Thoughtonomy in June, but collectively they are still worth 18 times more than they were post-IPO, including an issue of new stock.
What investors are concerned about is the company’s deepening losses. While 1H 2019 sales rose by 82% to £40.4 million ($51 million), they were nearly matched by an adjusted loss of £34 million ($44 million), which the board has dismissed as a by-product of aggressive expansion. Blue Prism is three times bigger than it was in the previous year and now employs over 1,000 people.
That said, its losses are over eight times deeper year-on-year, which is a cause for concern. But arguably, they are small beans compared with the nine- or ten-figure losses that Uber or Tesla, for example, have sustained for years. Yet everyone still raves about those companies in a US tech culture in which success is about being seen to succeed.
But if tech success is about making a big noise, the UK’s silence about Blue Prism has been deafening. Is this one reason for the widespread belief that – with a handful of exceptions – only the US, China, or India can launch a successful software company and grow it worldwide, without selling out to a US or Japanese giant? Even DeepMind didn’t manage that.
A challenge, then, is that – beyond its shareholders’ volatile support - which has seen market cap tumble in six months – Blue Prism has been expanding in a freezing vacuum in the UK. It’s hard to imagine the same media silence happening to a company in Silicon Valley, with its temperate tech-investment climate and the oxygen of relentless publicity.
New (old) blood
It may also be why there have been changes at the top in Blue Prism. Eric Verniaut joined as COO in August. Alastair Bathgate remains CEO, but with a renewed focus on customer relationships. And on 22 October it was announced that Kingdon was returning as Executive Chairman after time out in a non-executive role while he developed other projects – including Glass.ai, of which he is co-founder and investor.
Back in the hot seat at Blue Prism, Kingdon believes that RPA is on the cusp of being recognised as “the platform of choice for all corporate AI”. So diginomica pulled up a chair with him to discuss the challenges of growing a company that few people have heard of. He says:
RPA is, in my mind, really AI for business, and will be the channel through which this powerful, fast-evolving technology breaks through to mainstream businesses.
But who is Jason Kingdon? He’s a technologist and mathematician with a PhD in computer science from University College London (UCL), where he co-founded the Intelligent Systems Lab that went on to nurture DeepMind. He was one of the first people to deploy a neural network in live financial forecasting as far back as the early 90s.
While at UCL, he formed a company called Searchspace to develop intelligent transaction monitoring. Soon it had produced a system for the London Stock Exchange, the New York Stock Exchange, and others, but Kingdon sold out to a private equity group in 2005. What had started as a business with broad, ambitious aims had ended up with a narrow focus on anti-money-laundering (AML) technology for banks and the middleware to link it to other systems, he explained.
In search of a wider opportunity in 2007, he was introduced to Blue Prism – founded by Bathgate and CTO Dave Moss – and the relationship developed over the ensuing decade:
I thought they solved automation in such an elegant, light way, there's a real genius to it. They were looking through the telescope the right way, taking the user interface seriously and repurposing it as a machine interface. Focusing on the interface meant that you could theoretically connect it to everything from a Babbage machine to an Oracle system, or to something that hasn't even been released yet.
The lightbulb moment came when Kingdon was introduced to an early RPA adopter, a major online retailer:
That company said, ‘We're going to introduce you to our employee Kate – she runs the robots.’ This was 2007. Kate said, ‘OK, well the best thing to do is just give you an example. The last time we used them, it was because we had overpaid on 2,000 accounts and we had to reverse those payments. So I showed the robots how to change them on the accounts, make a reconciliation against the audit system, put an audit note against the way the system had managed the accounts, send letters out to individual customers telling them what had taken place, and put in a closing note.
All that took just a morning and they ran the robots that afternoon. So I asked, ‘What would you have done if you hadn’t got that capability?’ And she said, ‘Well, we would have had to take staff off the phones and off other tasks to sort it all out and it would probably have taken two weeks.’
So I asked what her background was and she told me she had left school at 16 and had originally worked in the company’s call centre. I said to her, ‘So what does it feel like to be a computer programmer now?” And she just brushed it off and said the task had been so easy to do.”
It’s a neat anecdote, but it gets to the heart of another challenge facing robotics and automation of every kind: what these technologies actually do – augmentation – versus what a hostile media believes they do - throw humans on the scrap heap.
Our conversation took place in the same week as yet another news story broke warning of mass job losses to automation and AI, this time coming via recruitment firm Robert Walters and market intelligence provider Vacancysoft (publicising a report that had been published a few weeks earlier).
That report warned that up to 30% of all UK jobs could fall to automation, especially those in middle-management. It quoted a 2018 World Economic Forum prediction that up to 75 million jobs could be lost to the world economy, many of them in once-safe white-collar positions.
Less often quoted is the WEF’s prediction that 133 million new jobs would be created as a direct result of the same Industry 4.0 innovations – a global net gain of 58 million. The nature of some jobs would change, but productivity increases and new opportunities, products, services, and companies would coalesce around the incoming technologies, just as they had done with mobility and ecommerce. That’s the message that isn’t getting through to a sensation-seeking mainstream media.
After all, if it were true that ‘robot in’ equals ‘humans out’ in such a simplistic way, then surely there would be a mass of evidence of this happening in the world’s most automated, robot-dense nations. Inconveniently for the doom-mongers, there isn’t – at least not yet.
Consider these figures. The most automated nation in the world is South Korea, which has a robot density of 631 for every 10,000 human workers. As of August 2019, unemployment in that country stood at 3.1% – significantly lower than the UK (3.9%), which has one of lowest robot densities in the developed world at 22nd in the automation league.
China aims to be in the automation Top 10 by 2020, yet it also has an unemployment rate lower than the UK’s: just 3.6 percent. Japan boasts 23% of all the robots on the planet and yet has just 2.4% human unemployment. The list goes on: Singapore (2.2% unemployment), Germany (3.1%), Taiwan (3.7%), and the US (3.7%) are all among the world’s Top 10 robot users.
In that 10, the exceptions in human unemployment terms are Italy (9.7%), Sweden (7.1%), Belgium (5.6%), and Denmark (4.8%). Those rates are higher than the UK’s, yet in most cases the jobless figures have been falling in what we can regard as the long tail of the 2008-09 recession.
So what about the US? According to PwC analysis of Bureau of Labor Statistics data, the most automated manufacturing sectors in America, including automotive and electronics, employ 20% more expert human engineers than comparable sectors – and pay them more. Many tens of thousands of new human jobs have been added to the US automotive sector since it purchased more than 60,000 industrial robots in 2016. So the idea that robots trigger mass unemployment is misleading.
Kingdon himself cites an MIT study of the automotive sector from 1970 to 2012, which found that although the sector is now fundamentally different, thanks to automation, it essentially employs the same number of people. Of course, RPA is about software, not machines in cages on factory floors, but the point stands.
Kingdon’s frustration at the misunderstandings over automation is all too apparent: it’s not about replacing human workers, he suggests, but about making them more efficient and freeing them up to support the business in other ways. But that’s not to say that no one will lose their job because of automation or that there are zero risks. The battleground will be transferable skills, education, and companies’ willingness to retrain their staff:
This whole media debate is wrongly informed by the idea that somehow there is a finite set of tasks and, once these things have gone, there's nothing left for anyone to do. In fact, it’s the complete inverse of that: we expand the economy, we expand the number of things that are out there to do and the number of tasks that can be done.
Another key aspect of robotics is that they mimic human beings. The reason they do that is the world is built for humans. So one of the things that gets missed in all this is that robots add things and drive consumption, because they consume things themselves. They actually grow the amount of economic activity, not reduce it. I don't think people understand that AI and robotics expand the economy.
But it’s not all plain sailing. Yet another challenge with Industry 4.0 technologies is that some organisations rush into adopting them for the wrong reasons: tactically rather than strategically, and to slash costs rather than make the organisation smarter or more agile or efficient. So it’s possible that workforces may be cut because of bad management and short-termism, not because of new technology.
So what next for Blue Prism as it moves into what Kingdon calls ‘phase two’ of its story – which hopefully involves telling it louder and more confidently? He posits:
I think that verticalisation is coming, along with the day when we have university-trained robots with deep sector knowledge. And to an extent that’s one of the conversations we're having with our customers.
We're trying to build a digital worker. I don't think there's anything more ambitious than that, with all its implications: how they work in teams, how they share. They can figure out issues with third party-applications themselves, avoid dead spots and areas that are running badly. They can self-optimise against the workload and what they need to do. That's the stuff we’re trying to build.
Connected RPA is another. We're trying to draw attention to the fact that this is a collaborative platform. That's very different from isolated robots sitting on people's desktops. But if you're going to create this vision, which we think is correct, that means organisations are going to have one-third people, one-third core IT, and one-third robots. That's where we think businesses are going.
As to where do humans fit into this world, Kingdon argues:
Design. How do you make your company better in terms of what you do? That's what human beings will be doing on one level. Another is empathy, dealing with human beings, dealing with human-to-human contact, creative work. We're a long way off people getting into an autonomous car that knows better than you where you need to go. Humans, I think, are going to take some beating.
Kingdon concluded our conversation by saying:
I think when the dust settles on this, people are going look back and say, ‘Wow, this is a technology that got invented in the UK. And no-one even noticed.’
Let’s hope they do notice – and soon.