Shaping the Future of Work

Rajeev Mittal, Managing Director – India, UiPath

Robot process automation (RPA) is about assigning repeatable tasks to machines, which over time acquire ever-greater sophistication. RPA requires human intervention only when the robot – which is really a set of software, and not a distinct physical entity – cannot handle a task, though the list of such ‘no-go’ areas will only shrink in the coming years. RPA has a whole range of applications, which will continue to grow. This will not enable people to do more meaningful, interesting work, and organisations to make huge efficiency gains.

Taking the robot out of humans…

Automating repeat processes with predictable outcomes

RPA currently works best in tasks that are repeated, high-volume and data-intensive, especially where the process is predetermined and the output is predictable. (Essentially, then, it automates repeat tasks that have fixed output.) This not only frees up the workforce from mundane duties, but also allows people to spend more time and energy on creative, strategic, higher value-add work. Agility and job satisfaction both improve. For the organisation, there are massive efficiency gains to be had, and in some cases, customer satisfaction receives a big leg-up. Importantly, compliance improved as humans’ ability to make mistakes is taken away.

…in multiple areas…

Everything from jobs and loans…

…to issuing credit cards…

…processing invoices…

…monitoring social media…

…and organising vast amounts of data

The uses of RPA are myriad, including several low-hanging fruit:

  • Job applications: robots can scan through social sites, identifying personality traits that the organisation might find desirable
  • Loan applications: RPA systems can ‘read’ an application, upload all the routine data (name, address, income), and check on credit ratings. Depending on defined parameters, it can provide an answer within seconds, with say, 90% confidence. Applications that fall outside a certain confidence band can be passed on to a human
  • Credit card issuances, claims settlements: Tasks that rely heavily on routine, back-end information can be automated, based on fixed parameters. This can greatly speed up processing time, raising customer satisfaction levels
  • Purchase ledgers: it is possible to have the system open an email, pull out an AP invoice, scan the details, input them into the ERP system, and even process the payment. The potential efficiency gains in FP&A are in the region of 95%.
  • Real-time social media monitoring: Bots can be trained to perform tasks like picking up on negative signals on platforms like Facebook, and taking appropriate action. Plainly, they can do this faster, and in greater volume, than a team of humans might
  • Organising unstructured data: 80-90% of all business data is unstructured, but advances in NLP (natural language processing), voice recognition, AI and machine learning are making it easier to capture and analyse such information. IBM’s Watson, for example, uses unstructured data, including x-rays, to arrive at probabilistic outcomes. In the next few years, the scope for such applications will expand dramatically

…and building a robot for every human

A digital assistant (or several) for each person on Earth

  • On the flip side, robots can – in time – become true ‘digital assistants’ to every human being. With cognitive automation, they will be able to start thinking like humans – setting appointments, managing calendars, even reading and – within fixed parameters – replying to emails. At UiPath, bots are used for everything from reminding workers to order their lunch and actually placing orders for them, to answering legal questions. They are, in other words, working side-by-side with humans, rather than necessarily replacing them.

Embracing the future

Plugging the missing parts of the platform…

From a technology point of view, the basic tools necessary for RPA have long existed, but some crucial nuts and bolts were missing. In recent years, advances in NLP, OCR (optical character recognition) and other technologies means that its scalability and range of applications has vastly grown. RPA is, however, more an ongoing journey than a fixed destination. It should also be noted that it works on the basis of probabilistic outcomes, rather than certainties. Moreover, any process that gets automated can only be as good or bad as the process itself – if you automate a poor process, the outcomes will be no better than if they were performed manually.

…and recognising it for what it is: one of many technological revolutions

Many worry that robots will replace humans, but as with any technological or industrial revolution, initial job losses will be followed by the creation of new, ancillary jobs. The automobile, for example, replaced horse-carriages, but it created a whole set of new industries: part suppliers, car dealers, service stations, and so on. As RPA becomes mainstream, the reskilling process will begin. RPA-related jobs that did not previously exist will get created, and whole organisations will spring up to service this need. More broadly, RPA will help build a more efficient ecosystem, one where the ‘robot’ gets taken out of each human in a very real sense, removing the mundane, improving efficiency, and freeing people up to do more interesting, value-adding things with their time.