Video: Robots challenge traditional economic analysis: Why raise wages?
If someone told you productivity in the UK is at the lowest level since the 18th century, how would you react? Skepticism would be warranted, because that's actually not true. Still, it provided the backdrop for an in-depth discussion on productivity and the future of work in relation to technology with one of the people in the know.
Guy Kirkwood is UiPath's chief evangelist. UiPath is among the leaders in Robotic Process Automation (RPA), with robots in this case being software robots configured to carry out repetitive tasks that humans previously handled. Think data entry, form submission, and the like.
Kirkwood, with a 20-year career in Business Process Outsourcing before joining UiPath, is known for making bold statements: "Outsourcing is dead" and "AI is bollocks" are some of those, and the "Fourth Industrial Revolution is coming" is the one that triggered this conversation.
More of that, please
Productivity is computed by dividing average output per period by the total costs incurred or resources (capital, energy, material, personnel) consumed in that period. So, to set the record straight, productivity in the UK today could not possibly be at the same level as in the 18th century.
The findings of research conducted by the Bank of England were different; what it said was that productivity growth in the UK is stagnating. Kirkwood sees this as a sign, and noted that today and the 18th century have one thing in common: They are transition periods.
"Today marks the period when post-industrial economies move into the fourth industrial revolution, one predicated on the automation of white collar work and the development and deployment of artificial intuition systems that will extend the cognitive power of an individual human," Kirkwood said.
Kirkwood expects that productivity gains wrought by the first industrial revolution will be repeated in this fourth industrial revolution, although it's going to take a few years to show.
"Looking at venture capital investment should give an idea. VCs look at market size, and RPA and intelligent automation cross-cut across all industries -- there's no limit to where they can spread. Soon, there will be one robot for each employee. The productivity gain will be massive," he said.
Kirkwood points toward his own experience to justify this claim: UiPath has raised more than 180 million in the last year or so, its client base has grown more than 10 times (from 100 to 1100+ globally), and it has either secured or is negotiating deals with the likes of Oracle, Microsoft, SAP, and Salesforce.
And it's not just UiPath. RPA as a whole is booming. The big questions are: Why, how, and what does this mean?
Kirkwood explained that RPA is the low-hanging fruit for automation, which can be applied everywhere:
"We signed the deal with Oracle in 2017, and UiPath will be natively built into Oracle's cloud business by the end of this summer. Robots will be there preconfigured for anyone who buys Oracle cloud services. You'll just need to activate them.
It's great for us to be partnering with Oracle, but I was recently discussing with a Forrester analyst and asked him why did he think Oracle did it -- what's in it for them. He said it's because it represents a shadow channel for them. This cross-cuts industries, and we sell directly to Line of Business, not to IT."
That may be an interesting sidenote in the cloud wars, but it's just one way RPA is spreading -- there's more. Most prominently, the big consulting companies. UiPath has partnerships in place with many of them, as they have established relationships with clients that are leveraged to introduce RPA.
"Every organization is going down the Digital Transformation path. Amazon and Uber have become verbs, nobody wants to have their business disrupted, and the ones that really get this are the big consulting firms. They're using RPA as the first step in that journey, because it's relatively quick and easy to implement, and the ROI is good.
They will go to organizations and tell them they will demonstrate value, starting with RPA. They will get the best people in the organization in a specific job, automate it, and then just show the board the difference between them and a robot doing the same job. The board will go, "Yes, more of that please."
Robots that just work
Unlike humans, robots work 24/7, so the gain in output is clear. But the system will work and productivity will increase only if the cost of automation is reasonable and quality of the result matches that of a human over time.
Read also: Research: Automation crucial for 60 percent of organizations | No, AI won't eat your job, say tech chiefs, and here's why | Business process automation: Where it works, and where it doesn't
Automating processes and integrating applications is not new. Tools for business process management and approaches for application integration have been around for a while. In fact, these are key aspects of what UiPath does. What's new, Kirkwood said, is the approach to capturing business processes initially and maintaining them over time.
UiPath's platform includes something called the Studio, a visual process designer used to capture processes to be automated. Although this approach has been around for a while, is well understood, and works, there are some issues with it.
First, it's expensive. It means that business and technical experts have to spend time to analyze, capture, and implement the process. As Kirkwood noted, this is always complicated, as people work in different ways, cut corners, and deviate from the official processes.
Second, it's brittle. Processes typically involve a number of systems, and every time something in any of those systems changes, the automated process will stop working and will have to be adjusted.
That's what's killed robots before, but Kirkwood said UiPath found a way to deal with that:
"With the advent of computer vision, we were able to automate over Citrix as well as across underlying apps. That's what's driven adoption in outsourcing countries.
In most cases, they don't have access to the applications, they just use them remotely. So we only have access to a picture, a bitmap. But we are able to scrape those screens with accuracy, and detect for example if a field changes, so we readjust and it does not break the robot."
Kirkwood goes on to add that there is an additional element that makes integration hard -- proliferation of applications from many vendors:
"There was this saying, nobody got fired for buying IBM. Now, it happens, and you also see it in the valuations of these big companies, because the days of big monolithic companies doing everything are gone.
Organizations figured out they can use a combination of tools to get what they want, plus many organizations have grown through mergers and acquisitions. In automation, that includes things like process mining. We work with our partner Celonis there. We monitor what a human does, Celonis maps this to a process, and the outcome is a robot."
As for the brittleness issue, Kirkwood mentions more partnerships in UiPath's ecosystem that help deal with that: "We work with ABBYY, which turns semi-structured data to structured data. The data will be there, maybe not in the same place, but ABBYY handles that. Enate improves workflow between humans and robots. Typically, there are parts of the process that humans handle better, and Enate manages the handoff. And Humley uses industry specific ontologies to handle context and helps us work with voice."
UiPath employs two types of robots, front office and back office: "The back office robots do the shitty jobs nobody wants, and they do it unattended, with checks and balances in place to enable them to adjust almost automatically. Front office robots are attended; for example, getting all your information and serving it to an agent handling your call," Kirkwood said.
The joy of work, and the 'Jen problem'
Kirkwood noted that it's a conglomeration of technologies that makes this possible. He also reports on success stories for UiPath, which allude to a broader issue.
Read also: Why it's time to prepare for a world where machines can do your job | When robots eliminate jobs, humans will find better things to do | Executive's guide to AI and the future of IT jobs (free ebook)
Hollard, a South African insurance company, was able to handle 98 percent of its 1.5 million emails autonomously, lowering the cost of transactions by 91 percent and achieving a six-times increase in speed.
SMBC and Dentsu in Japan are among UiPath's bigger clients, and Kirkwood said for them it's a way to deal with the Japanese culture of overwork. In India, KPMG automated 45 percent of its audit processes using UiPath, which Kirkwood said achieved higher accuracy with lower cost than the previous effort using IBM Watson.
Adoption of AI is slower than RPA, which ranges between 12 percent and 18 percent, Kirkwood, said and we should expect to see this spread like wildfire over the next 18 months and beyond. Ultimately, he said, RPA will disappear, in the same way water and electricity have: It will be there and just work, and nobody will care exactly how, as long as it's there at the press of a button.
"The underlying infrastructure of organizations will be technology rather than people," he said.
"RPA is at the heart of this." But what will happen to people then?
Kirkwood said that even though organizations typically approach RPA with the intention of automating to reduce their headcount, it turns out differently:
"They all do it for the same reason, but it's the wrong reason. What they find is the value of their people once they take away the shitty jobs they have to do. It's not just customer experience that's improved, but also employee engagement. Organizations are reusing people in much more effective and efficient ways, and the culture changes.
The word Joy was actually used to describe this. A spokesperson from Generali's Irish operation recently remarked in one of our events that the mood has changed, people are happier, and service is now measured in terms of compliments."
Gartner recently forecasted that by 2020 automation will create more jobs than it destroys. Kirkwood seconds that, adding that his experience is that people are freed to engage in more creative activities as a result of introducing automation, new services and products can be created, and the economy grows.
He also noted there can be reduction in headcount, but that mostly happens through the following:
"Natural attrition, not layoffs. Some are able to adjust and move up the value chain, some are not.
Right now, India represents 9 percent of our revenue. And there's 3.5 million people in India doing tasks that will be disrupted, or eliminated, by automation -- the same tasks that are currently offshored. But at the same time, I see that 60 percent of the 55 thousand people that have attended courses in our free MOOC platform are from India.
This means that although top-down change may be slow, bottom-up change is fast. People are bringing themselves up to speed to adapt. But then there's also the "Jen problem."
Jen is someone who has been in the organization for 30 years. She does accounts payable on one day, accounts receivable on another day, and maybe some HR. From an outsourcing or automation perspective, this is a nightmare. How do you outsource Jen? And after processes have been automated, there's very little she can do.
What can we do with the Jens? There's no clear answer to that. Lifelong learning, universal basic income, tax on robots... The thing is, automation will become ubiquitous, and we as vendors have to think about the impact on clients and people. That's unusual."
To circle back to productivity, there is an ongoing debate as to whether workers have received a raw deal out of the increase in productivity made possible by technological progress. Kirkwood may offer an optimistic view of the future, one in which workers are liberated from the drudgery to joyfully engage in creative tasks, but one has to wonder:
Does efficiency equal work nirvana? Is the world becoming more decentralized, or more centralized? Is the concentration of power and control in the hands of the Amazons of the world, in part due to their technologically-boosted efficiency, a good thing for people and for democracies?
More interesting and creative work is an improvement. But perhaps more important for the well-being of workers is the role of the power structure within firms. Time will tell whether the change of culture Kirkwood sees will extend to the point of acknowledging firms as entities operating on a democratic basis.
- How to automate the enterprise: Your guide to getting started
- Research: Automation of low-level IT tasks is already widespread
- How robots are filling worker shortages, replacing 'bad' jobs, and making work more rewarding
- How to do automation right
- IT automation: Where it's working, where it's not
- Five best practices for bringing automation to your organization
- Chat bots: How talking to your apps became the next big thi