When you started a new role at your company, you probably had someone explain the company’s products, services, and processes, and go into detail about your role and responsibilities, writes Neil Murphy, Global VP at ABBYY.
You may have even received a handbook detailing this information. This onboarding process was crucial to ensure success in your role, and so you’re able to perform your duties. Unfortunately, this same roadmap to success isn’t being applied to the new digital workforce.
The number of digital workers entering the workforce will increase by 50 percent by 2021, according to new IDC research. With the explosion of the robotic process automation market (RPA), there are now millions of digital workers employed at businesses around the world. So it shouldn’t be surprising that we will soon see organisations give at least one robot to every employee to augment their day-to-day activities.
But despite the promise that trillions of dollars are expected to be saved by deploying digital workers, most RPA projects fail to fully deliver on that promise. The root cause of many of these failures is that digital workers don’t know how to handle unstructured content or undocumented processes – just like badly onboarded employees don’t. In order to realise the return on investment in RPA and content intelligence, we need to ensure our robot colleagues are employed and onboarded appropriately.
Here’s a guide to employing a digital worker (for the not-so-digital worker).
Step 1: The Interview Stage
The first step is determining if you really should hire that digital worker. Given all the hype this may seem counter-intuitive, but not every process is qualified for RPA. Even worse, picking the wrong process will only lead to frustration as you try to make your digital worker perform a task it technically is unable to do.
To determine whether your process needs digital workers, you must ensure it follows rules-based decisions rather than judgement-based. It is a robot, after all. If your process is prone to human error and is repetitive – and especially if there is input data, and it is digitised through OCR and document capture – then the opportunity is ripe.
Step 2: The Job Offer
Congratulations on your new employee! But don’t get too excited. Before they settle in you’ve got a few housekeeping tasks. Ensure you avoid duplicate work and overlaps in the job function for your digital worker. Unlike their human counterpart who can speak up when something goes wrong, digital workers will tirelessly do what you ask of them, even when there are unintended results. This makes it all the more critical that digital workers’ ‘managers’ use process intelligence to ensure they are properly designed to avoid conflicts, and can deliver their benefits without the costly side effects of a poorly employed worker.
Further to this, the success of our human colleagues is dependent on documenting the process so they know what to do. For a digital worker, it is even more critical that a process is fully documented, as this information is the basis of properly training new digital workers.
Having this insight into processes means you can evaluate your current processes in their ‘base line’ state, so that process automation teams can clearly set ROI expectations, and ensure agile service delivery and that automation efforts do not produce any unintended consequences.
Step 3: Training the Technologically Advanced
As with our human workforce, if we want our digital workers to handle increasing complexity and process sophistication, they will need more training. However, there is no skills gap with digital workers. This training is realised through the addition of cognitive skills such as AI and machine learning enabled content intelligence, to raise the digital IQ of digital workers. Only then can they understand, reason and learn on a continuous basis.
Previously, first generation RPA bots focused on automating high-volume, relatively simple processes involving structured data with no human intervention. As enterprise demands have evolved and AI capabilities increased, digital workers are increasingly being used in processes with unstructured data, in more complex environments where humans are part of the process, and where some cognitive reasoning may be needed.
A prime example is in the automation of invoices. Digital workers are trained on a company’s ERP system and a handful of invoices for its course intelligence. Then, their content IQ continuously increases by monitoring and learning from variations in invoice forms, data and how exceptions are handled. Eventually, they can achieve straight-through processing without any human intervention.
Step 4: Review Meetings
Human workers regularly receive performance evaluations, and it’s important for digital workers’ performance to also be monitored and corrected. It’s another common reason why RPA projects fail – bots aren’t monitored effectively and get stuck performing broken or poorly executed processes. Automating a bad process just makes bad things happen faster.
Using process intelligence to monitor digital workers ensures your automation investment is operating as expected post-deployment, especially in mixed mode scenarios where bots incorporate human assistance. Beyond digital worker monitoring, process intelligence can easily specify detailed scenarios or conditions that trigger real time alerts to the right people at the right time – to handle a bottleneck immediately or directly trigger a remediation task or process in another technology.
Finally, process automation leaders can ensure the promise of economic gains are realised by calculating clear, quantifiable post-implementation cost impact that is monitored daily. This provides data-backed justification for future automation initiatives.
Step 5: Rewarding Robots
By having proof of digital workers’ performance and cost impact, you’ll be able to give bots a “promotion” and expand them at enterprise scale. However, scaling from tens to hundreds, or even hundreds to thousands of bots requires significant command and control to ensure automation remains synchronised across every process and business system it touches. You need a central viewpoint to monitor all bots and the contributory role they play in business processes, even those that cross different business silos, which process intelligence helps achieve.
The future of work will be made up of a growing digital workforce that will take on more reasoning and decision-making, allowing them to go much further than simple automation. IDC estimates that today machines conduct 29 percent of evaluating information, reasoning and decision making, and that this only increase from this point on. It’s important that we are prepared for this new class of workers and use process intelligence to create their path to success, just as we use all our people skills with traditional employees. Being able to manage both humans and robots is the next big skills challenge – now is the time to learn.