Demystifying automation technologies

    New technologies such as robotic process automation and artificial intelligence offer potentially huge benefits to forward-looking business services companies. Automation expert Andrew Burgess, Noel Lyons, Barclays director of digital design, and relationship director Lee Atkinson explain some of the current thinking.

    An artificial intelligence system recently taught itself to play space invaders in a project run by Google-owned company Deep Mind. Its success was another landmark in the development of artificial intelligence (AI), which, along with robotic process automation (RPA), is delivering huge value to many companies.

    RPA involves software robots that work across a company’s systems replicating the kind of repeatable, rules-based processes often carried out by humans.

    Andrew Burgess, a strategy adviser and consultant in the areas of robotic automation and artificial intelligence, explains that new customers used to wait 24 hours or more to have their number switched by one mobile phone provider. Now it can be done online in 10 minutes, with a robot working in the background. And a major courier company cut the length of its claims process from six weeks to a few minutes by introducing RPA.

    Burgess said that using a software robot can cost about a third of a traditional offshore business process outsourcing person. Compared with onshore outsourcing, the costs can be as low as one-ninth.

    Self-learning systems

    “While RPA is designed to replicate processes carried out by human hands, AI is all about the brain and has self-learning capabilities,” said Burgess.

    Some of the main AI capabilities are currently image and voice recognition, search, and data analysis (or ‘clustering’) of data. So, while the robots need structured data, AI systems can process semi-structured or unstructured information ranging from images and voices to documents like invoices, which are rarely identical. AI systems can decipher natural language, employ reasoning and problem solving, and predict what will happen next.

    The power of AI image recognition, for example, can be seen in the FindFace app, based on the huge library of images on, Russia’s Facebook equivalent on which profile pictures are mandatory. Using these pictures, it can identify the names of Russians with 70% accuracy simply by pointing a smartphone camera at them.

    But there are far more serious applications of AI. One provider of insurance services to carriers uses AI to read insurance documents, which vary from company to company, while ‘magic circle’ law firms use the technology to read through contracts.

    Burgess added: “PayPal is particularly good at using predictive data to identify fraudulent transactions. You might think that’s been going on for years, so that you might get a message from your credit card company saying ‘were you in Bolivia at this time and did you buy that’. PayPal has taken the technology a step further, so almost in real time they can identify fraudulent transactions. The industry standard success rate for this sort of fraud is 1.32% but PayPal has reduced it to 0.32%.”

    A decade ago, one large investment bank, Goldman Sachs, had around 600 equity traders based in its New York office. Now it has just two, supported by 200 software engineers. And Virgin Trains has combined AI and automation systems so that if you email the company your message may be read by an AI system. All the information is gathered, validated and forwarded to the right person, or the right robot.

    Reducing costs

    The changes that RPA and AI have brought so far are just the tip of the iceberg and society will have to adapt as the pace of change increases. For example, there is the obvious impact on employment. Giant tech companies embracing these technologies now have vast market capitalisations with relatively few employees.

    But before rushing to introduce this new technology it’s important for businesses to understand how it can help them. It may not be right for all tasks and, particularly in heavily regulated industries, it can present a risk.

    Burgess believes that the key question for businesses shouldn’t be: how can we use RPA and AI in our business? It should be: what are our business objectives, and how can automation help to deliver them?

    When Burgess works with a company he develops what he calls an ‘automation heat map’, by looking at the different technologies available and how they might apply to different areas of the business. With that he is able to look at the possible benefits offered by the technologies and the company’s priorities so it can develop a roadmap for the introduction of RPA and AI.

    The impact on banking

    So, how are these technologies affecting the world of banking?

    Barclays recently launched its own AI-based digital assistant with the aim of helping customers stay financially fit. Noel Lyons, director of digital design at Barclays, explains that inspiration for the project was drawn from fitness apps like Fitbit. The first iteration was tested on a small number of people who liked the concept but wanted it to do more.

    Lyons explained: “In June 2017 we rolled out the first proper iteration of our financial coaching system. It will give you genuine insights and breakdowns of your spending. It’s out with 25,000 customers, they’re using it and we’re getting the data back. We’re trying it out with people, finding out how they’re using it, if it makes sense, whether it’s helpful over a long period of time and how it’s changed their behaviour.

    “One of our big learnings was that this kind of service could allow us to build a relationship over time. So, imagine it’s payday and we change the image to a happy one and start talking to the customer about saving. Talking about saving on payday is much better than talking about it a week before payday.”

    Everything from colours to tone of voice and how the system was received by customers in their 70s has been tested. A major result has been that the AI system appears to be three times more likely to convert people into customers than the existing system and various media companies have reported similar results.

    This kind of technology could complement existing service levels. For example, it could be used to carry out the functional tasks in call centres, freeing up staff to deal with more complex issues for customers, where real human interaction is needed.

    Business implications

    The infiltration of RPA and AI into business strategies over the past few years has led many UK companies to rethink the way they will do business in the future. These new automation technologies are helping businesses to streamline processes and customer journeys, manage costs and improve business development. Those companies that don’t look to embrace RPA and AI in appropriate areas of their businesses risk getting left behind.

    Lee Atkinson, a relationship director at Barclays who works with IT companies, sees a number of challenges for firms looking to navigate a successful AI journey.

    “Directors are grappling with AI strategy, they are often unsure how AI can be utilised affectively within their business in a cost-effective manner. In the early adoption stage its usually unclear whether AI will work and what the benefits will be. Often the initial investment can be off-putting, there are businesses that allow you to investigate how AI works best for them through proof-of-concept projects that run alongside existing embedded processes before looking to adopt this.

    “Finding the right skills can also be a challenge. Data scientists tend to be expensive and in short supply.”

    Lee Atkinson
    Lee Atkinson

    To find out more about how we can help your business success contact:

    Lee Atkinson

    07775 542589

    The views expressed in this article are the views of the author alone and do not necessarily reflect the views of the Barclays Bank PLC Group nor should they be taken as statements of policy or intent of the Barclays Bank PLC Group. The Barclays Bank PLC Group takes no responsibility for the veracity of information contained in the third party guides or articles and no warranties or undertakings of any kind, whether express or implied, regarding the accuracy or completeness of the information given. The Barclays Bank PLC Group takes no liability for the impact of any decisions made based on information contained and views expressed.

    Barclays Bank PLC is registered in England (Company No. 1026167) with its registered office at 1 Churchill Place, London E14 5HP. Barclays Bank PLC is authorised by the Prudential Regulation Authority, and regulated by the Financial Conduct Authority (Financial Services Register No. 122702) and the Prudential Regulation Authority. Barclays is a trading name and trade mark of Barclays PLC and its subsidiaries.