With advances in machine learning, artificial intelligence and big data, companies can enhance their ability to predict rather than react to rapidly changing demands and expectations.
By implementing a digital workforce of software robots, organisations can ensure that work is done around the clock, eliminate human error and reduce human dependency to drive revenue and ensure an ‘always-on’ service for customers.
Intelligent process automation (IPA) is improving results in many sophisticated processes such as loan applications for banking, claims adjudication for insurers, provider verification for healthcare and clinical data management for life sciences, as well as traditional technology processes like infrastructure services and information management.
However, the fact remains that automation still has its limits, and there are some things that robots just cannot do, such as medical management, underwriting, case reviews, speak or comprehend colloquial slang, understand people’s emotions and think on their feet.
It is a stretch to say that robots and software now run industry. Yet it is fair to say that process automation – working alongside smart people, rather than supplanting them – plays an increasingly important role in how organisations operate.
‘Despite all the fearmongering around the issue, the rise of robotic process automation (RPA) services in the workplace is actually liberating people from unproductive work and creating whole new industries and opportunities,’ says Gajen Kandiah, executive VP of business process services at Cognizant.
‘Businesses that are embracing these technologies are capturing more data, improving processes and generally empowering workers to be more effective at their jobs, thereby improving the customer experience.’
RPA was introduced specifically to automate routine and mundane tasks, with a view to freeing up employee bandwidth to allow workers to focus on core business objectives. By eliminating the need for humans to perform certain manual duties, it facilitates opportunities for those staff to be creative, rather than taking jobs away from human workers.
It is for this reason that businesses across multiple industries are using RPA as a vital component of a broader digital transformation strategy. Business process management (BPM) providers, for example, are increasingly implementing it for high-volume and repeatable processes such as transaction processing and data entry, freeing workers to think creatively and use their time to focus on core business goals.
In terms of variation, RPA has a huge number of applications and can help to simplify data gathering and analysis, enhance flexibility, improve compliance by providing detailed audit logs, and enable 24-hour flexibility that could not be achieved by a single manual worker.
‘Implementing RPA can go a long way in reducing costs and has the potential to save millions if used the right way,’ says Keshav Murugesh, CEO at WNS. ‘Ultimately, business is about driving revenue to improve the bottom line, and automation allows workers to focus on their core business processes that drive revenue.
‘The organisations that are best at using RPA will not see it as a process independent of human intervention, but one with many opportunities for humans to work alongside robots.’
The term ‘cobotics’ has been used a lot in recent times, highlighting how important the human input into automated processes really is. Human judgement is of paramount importance when it comes to the implementation of RPA and only reinforces the notion that the purpose of automotive processes is to streamline human labour and not replace it.
As we move towards a more connected, technology-oriented world, RPA is likely to play a much more prominent role, with businesses devoting significant chunks of their budget to it. Real-world robots have a vast number of enterprise applications, but businesses must appease the fears around a robot workforce revolution. ‘They must identify how RPA can optimise the work of employees further,’ says Murugesh, ‘rather than how it might be taking away their jobs.’
When examining the specific technologies that enable business automation, data analytics and machine learning are quickly identified as key innovations. However, it is important to differentiate between the type of analytics and the ties between them. There are two main types: predictive and prescriptive.
Predictive analytics and associated tools are key to business automation, as it can utilise both historical and recent data sets to gauge the probability of outcomes while highlighting uncertainties associated with the methodologies used in forecasting.
This enables a human or a machine to make informed decisions to enable (or prevent) a future outcome. Prescriptive analytics then codifies these suggested actions in the form or rules, policies or constraints.
Machine learning, however, will take data analytics a step further by providing cognitive computing capabilities to learn and make human-like decisions that continue to improve, and to identify trends.
‘Let us assume that prescriptive analytics finds that a type of computer virus is on the rise,’ says Joe Kim, senior VP and global CTO at SolarWinds. ‘Machine learning can then identify this trend and utilise the forecasted outcome and associated uncertainties to create predictive rules to ensure that the virus does not further impact the environment by automatically patching machines with up-to-date software.’
The possibilities with speech analytics are endless, and companies are just scratching the surface of how this technology can be used to enhance the customer experience and reduce employee churn.
One growing area of interest is how these AI engines detect emotion and sarcasm – giving brands better insight into what customers actually feel.
There is a significant amount of customer sentiment buried within word usage and sentence structure. With a clear understanding of customer sentiment, brands can evolve sales, marketing and customer service functions to improve customer engagement.
‘As more data is collected, businesses can apply that language expertise to larger data sets, and subsequently develop increasingly accurate predictive models,’ says Brad Snedeker, director of innovation at Calabrio. ‘The sooner we can predict customer behaviour accurately, the sooner sales and marketing can take action to cultivate loyal, long-term customer relationships.’
In the accounts payable department, data entry has long been associated with automating repetitive, high-volume business transactions, such as invoice automation and remittance automation.
The rise of automation technologies is helping tackle the long-held and archaic problems associated with manual-driven travel and expense management. Development of intelligent mobile receipt processing technology, operating in real time to execute ad hoc transactions, has the potential to ease processes for both employer and employee.
In the legal department, more advanced information extraction scenarios are using intelligent technologies to navigate and explore more complex, unstructured document sets, such as contracts, claims, customer case files and corporate investigations.
The impact is twofold: the ability to ‘speed read’ complex language documents, thereby significantly improving the speed of approvals and investigations, and the ability to intelligently extract key data points and key insights that might otherwise be missed by humans, due to the sheer amount of information that now needs to be processed.
When facing even larger data sets, this problem and opportunity is magnified. Varying degrees of classification and extraction are also being applied to power a number of self-service systems.
These can be simplistic chatbots that work based on keyword analysis, or more intelligent customer services agents that have a much deeper understanding of customer questions and are then able to respond accordingly.
‘Thanks to automation, a high rate of incoming customer requests or complaints can be dealt with in real time, significantly boosting customer service levels,’ says Jonathan Darbey, product group head at ABBYY. ‘More complex queries that need a deeper personal touch can then be handled by staff, who are no longer time-pressured. The result is win-win.’
As routine work is becoming automated in contact centres, speculation about what this means for certain sales, marketing and customer service roles continues to grow.
It seems that people skip a step in determining how this evolution will affect them. With call centres, for example, there’s no desire to replace the human element, largely because people engage with each other for emotional reasons, and customers want a human for help and counsel.
‘Certain roles within the business need to upskill and take on new roles now that the data has moved beyond human processing capability,’ says Snedeker. ‘But despite the creation of roles focused on what we do with insights rather than raw data, it seems that the human resistance of AI and automation will remain a slowing factor.’
When it comes to specific industries, there are several that stand out as being more susceptible to automation in the coming years.
In the fast food industry, it’s already happening. Today, many fast food chains have kiosks that allow customers to make an order and then go to another queue to collect their food. There are also fast food chains in the US that allow you to order your food online and then go to the restaurant to pick up your order when it’s ready.
Another industry is IT support. Automation will be used for tasks such as raising tickets and even first-line support and resolution suggestion, meaning that there will be fewer people needed on call desks. Automation will lead to selective sourcing and will allow companies to become more resourceful and more cost-efficient.
Hospitality will also be impacted. ‘On airlines, automation already exists in self-service check-in desks,’ says Kim. ‘We will start to see this replicated in the hospitality industry, with hotels offering similar services that allow guests to check in and out of rooms themselves.’
Businesses must manage the culture shift required to makes sure that transitions are successful. Every significant, well-publicised report on automation indicates the impact on the existing workforce as a key threat to society, and it has been well advertised to all workforces that using automation tools could be the beginning of the end for their existing roles.
Managing that perception and keeping a workforce enthused and motivated while still taking advantage of newer technologies will be a key challenge. Failing to do this can be a significant barrier and deterrent to the adoption of automation.
Any automation project will require significant engagement from the personnel carrying out a task targeted for automation to ensure that the processes and workflow are captured accurately and automated efficiently.
‘A resistant workforce risks piecemeal automation, requiring unnecessary levels of workforce interaction and undermining the overall potential of the automation project, resulting in reduced efficiency and cost savings,’ says Andrew Joint, commercial technology partner at Kemp Little.
But while machine learning enables departments and business functions to be proactive and mitigate risks, it doesn’t mean that a human element isn’t still needed. Tribal knowledge – the knowledge gained from experience – is still essential for business decisions to be made in suitable context.
In essence, no matter what thought processes are being devolved, a person will need to decide whether they want an automation trigger set based on a particular rule learned.
For example, within an IT environment, an AI tool might decide that CPU utilisation on a server is insufficient to maintain regular system performance. As a result, the AI may decide that a trigger must be set, as low CPU power may result in load issues, or could be an indicator of a virus or malware.
But a person in the business might know that this peak is an anticipated occurrence, and is short-term so, in fact, no trigger is necessary.
‘While the data and “thinking” that AI and automation is capable of is valuable, the human element is needed to enrich the conclusions reached and ensure the correct course of action is taken,’ says Kim.
The reality of automation usage is that businesses are only just starting to fully implement automation solutions, and even then it is typically in ‘ring-fenced’ areas – not entirely across an enterprise.
Early usage of automation in an organisation tends to focus on more low-level data entry automation. But once the benefits are seen, organisations look to automate more ambitious and complex areas, where ‘intelligent’ or AI solutions are more suited to making reasoned decisions.
The rewards in this area can be significant, but so can the risks if not properly managed. As the expectations of what automation can do continue to increase, it will become more synonymous with artificial intelligence, rather than the more process-driven automation solutions seen in the past couple of years.
‘Like other technologies that have come before it, I think 2017 will see the use of automation tools move along the “diffusion of innovation” cycle beyond only the innovators to the early adopters and early majority,’ says Joint. ‘I think we can also look at the development of other technologies like cloud computing – where the stack of integrated offerings was the natural next stage to individual SaaS/PaaS/IaaS offerings – to understand how things may develop.
‘The integration of a number of different automation tools is the likely next phase of development, with an increasing focus on AI, and we may begin to see clearer plans for that in 2017.’