This article was written by Roger Camrass, director of CIONET UK and a visiting professor of the University of Surrey, and is based on conversations during dinners on ‘Intelligent Automation’ sponsored by Blue Prism, Accenture and Deloitte in London this November.
Much hype has been generated over Robotic Process Automation (RPA) and Artificial Intelligence (AI) in recent months. The prospect of a ‘Digital Workforce’ is on everybody’s lips. But just how real is this prospect, and what forces might bring RPA and AI into common use? So far, expectation and reality remain far apart. Dinner conversations amongst the UK’s largest companies shed some valuable light on the pace and direction of take up.
What is perhaps the most interesting insight from such conversations is that AI could be driven by external, consumer driven forces rather than internal corporate endeavours. Almost two years ago the CEO of NVDIA announced that his AI chips were being adopted by leading white goods and car manufacturers across the Globe. At the same time Digital Natives such as Google, Apple, Amazon have been buying up AI companies.
Our view is that such external developments could drive the pace of AI development within large incumbent organisations rather than internal programmes focused on productivity and cost savings.
A confusion over definitions
Delegates to both our recent dinners appeared to be confused about the distinction between RPA and AI.
RPA is a term that was invented by the CTO of Blue Prism, a global leader in this technology. It applies to the automation of simple, predictable processes by applying rule-based procedures. Benefits of RPA programmes include productivity improvements, cost savings and enhanced customer experiences. RPA is often associated with BOTs – software programmes.
AI applies to more complex, variable processes. Despite being 50 years in its inception, AI has only recently broken away from rule-based systems to apply learning algorithms that constantly adapt to external conditions. Digital Natives such as Google have acquired AI companies such as Deep Minds to help make sense of the myriad of data available on the Internet.
AI heralds a new era of personalisation and computer intimacy.
What is the vision for Intelligent Automation?
According to IBM we have generated 90% of all data in just the last two years. The challenge today is to convert such data into useable information, or knowledge, from which valuable decisions can be made. In our digital economy ‘information about a transaction can often be more valuable than the transaction itself’. Intelligent Automation can extract commercial value from the many different transaction points that define our lives as consumers – from finance and healthcare to food choice and entertainment. AI has matured to the point now where such value can be extracted.
At the same time, we as managers recognise the waste of human resources in many areas of our working life. We just need to examine how we spend time daily to know that automation could give back hours of valuable time. Public servants such as doctors, police, teachers frequently spend up to 50% of their time on administration rather than performing value-adding tasks. Automation could help double the capacity of our NHS without any additional manpower.
How is automation being applied today?
There are many good examples of how Intelligent Automation is generating commercial benefits today. Utilities are applying automation tools to match demand and supply in real time to optimise resources (power stations) and increase profitability. The prospect of distributed power (localised generation through solar energy and wind) will require such tools to balance national power sources.
In healthcare, automated receptions have been introduced in over 100 hospitals in the UK, reducing queues and improving the accuracy of data input. AI is being applied to electronic patient records to identify suitable candidates for clinical trials. This is likely to reduce such trials by months and years as patient gnome records become more widely available.
RPA is being employed in most UK call centres to deal with routine inquiries and to help streamline customer interactions. Here we see BOTs taking over conversations with the occasional human intervention. Natural language recognition is also helping to connect customers directly to computers in an effortless and timely manner.
What are the obstacles to Automation?
According to a recent survey by Deloitte ‘The Robots are Waiting’, most organisations have not yet reached scale in RPA and AI programmes. Typically, scale is achieved with over 50 BOTs, and few have passed this inflection point. Several potential barriers emerged from conversation including:
· Lack of IT support for RPA and AI programmes. Most organisations do not have the necessary skills in these areas to support growing business demand
· Regulation and compliance such as GDPR have eaten up much of the available IT resource in recent years, leaving little capacity to deal with such innovations
· Data is not in a fit state to support RPA and AI applications. Most data today remains the product of legacy systems and requires cleaning before automation can proceed
· Processes are fragmented. Few organisations have achieved necessary levels of standardisation across their many front and back office processes.
But perhaps the widest challenge is the lack of solid business cases to support large scale automation programmes. Typical implementation times for RPA extend between 4 to 24 weeks. Benefits rarely exceed $50,000 per implementation. This does not excite business executives.
How might external factors accelerate adoption?
If corporates appear slow to adopt automation the same is not true of their customers, especially consumers. Our homes, gadgets, cars are becoming equipped with high levels of intelligence thanks to embedded AI chips, as are our interactions with leading b2c companies such as Apple, Amazon, Google, Facebook and Netflix. Those participating in such supply chains – from retailers to manufacturers and financial services companies, will need to engage with consumer-led AI if they are to stay in the game.
The advent of IoT will further advance the prospects of intelligent automation within buildings, cars, cities, highways and utility networks. Within 5-10 years everything we touch or interact with will be intelligent. This will place even greater demands on those companies that operate in sectors such as property, construction, distribution and logistics.
Raising Automation to a strategic level
Automation at scale within the enterprise requires full ‘C’ Suite support. A case needs to be made for end-to-end process automation rather than local, tactical solutions. This involves multiple decision makers focusing on cross-functional automation rather than tasks. One important starting point here is the customer journey that travels across multiple regions of a large enterprise.
Greater resource will be needed both within IT and the business units to prepare the ground for automation, including the adoption of standard processes and the cleansing of data. Much of the tooling is now available within cloud platforms such as AWS and AZURE. The challenge is closer to people and process than it is with technology.
CIOs will need to work with their peers to develop broader automation strategies and be prepared to debate these issues with the workforce. Few staff see the merits of such automation today, fearing wholesale reductions in workforce numbers. Much work will be required to convince people of the benefits of a ‘digital workforce’ where humans and machines coexist together.
Questions yet to be answered
CIOs should help educate the ‘C’ Suite to resolve some of the current issues that impede broader automation:
· How to elevate automation (RPA and AI) to a strategic level
· What measures will be needed to scale up automation projects
· How to confront social and political issues within the workforce
· What might be the impact of external factors such as embedded AI
For more information please visit us at CIONET-UK