Those Robots Coming for Jobs? They’ve Already Arrived
A new batch of ominous warnings about job-killing robots just arrived from the usual suspects, very clever, credentialed, accomplished people mingling at a glamourous watering-hole: in this case, the Milken Institute's Global Conference in Beverly Hills, California.
These warnings tend to follow a predictable script: sufficiently urgent to get attention, but fuzzy enough to stop short of alarming.
Warnings about the impact of robotics on people’s jobs almost always have a temporizing quality: for example, at this conference robots are capable of taking over low-skill (document management) or (Wall Street spreadsheet jockeys) specialized jobs, yet aren’t at the point of actually eliminating broad job categories.
Which is puzzling: because it simply isn’t true. Robotic process automation isn’t at the tipping point of erasing large numbers of jobs: ironically, it’s already at work.
Perhaps the disconnect between how RPA is viewed at this conference ("Anyone whose job is moving data from one spreadsheet to another ..., that's what is going to get automated,") and what it’s actually capable of doing is simply a problem of perception: not so different from how a young adult coming home from college is often still seen as a child by his parents. It’s not hard to see why that could be the case.
After all, screen-scraping, later supplanted by web scraping, was the first way in which this technology’s ability to integrate at the UI level was used.
There’s no question screen-scraping is associated with ancient technology, and rightly so. It is old school, beginning as a one-way transfer of data from computer green screens via the terminal’s memory.
It’s also often an interface of last resort for an old school problem, the challenges organizations face in linking legacy systems with new applications, including lack of legacy documentation, source code and lost skillsets.
Web-scraping, on the other hand, isn’t ancient technology. It incorporates a cross-platform, open language document object model (DOM) to access and interact with HTML, XHTML, and XML objects. One illustration of web-scraping: companies employ it to automatically pull competitor’s prices from their websites, and run comparisons against their own. This use is what many people are referring to when they speak about swivel-chair automation.
But moving to web-scraping technology still left RPA with two major limitations, both of which reinforced the perception this robotic technology could only play a limited role, with limited scope, in business process automation.
Limited Role in Process Automation
While web-scraping is highly effective, it customarily takes place at the task level while employees typically operate at both the task and activity level. For example, an employee could be responsible for a pricing analysis activity comprised of several tasks: price comparisons; a summary report; and a price revision recommendation.
While automating one task, price comparisons, makes the employee more productive, it doesn’t eliminate the job he holds for the entire activity. It also limits the 24/7 capability of robots to the working hours of the person it assists.
Limited Scope in Process Automation
While robotic process automation software has been very effective at integrating on the GUI layer with web-based or internal systems, it has found it challenging to do with ERP systems running in virtualized environments.
The challenge lies in the fact that GUI integration relies on access to the logical elements of the user interface, but virtual technology – whether Citrix XenApp, VMware or Microsoft Hyper-V – keeps them isolated on a server behind a secure firewall: what appears on the user interface is actually screenshots of the ERP application running behind the firewall.
It was possible for robotic automation software to meet this virtualization challenge, but solutions were generally complex, time-consuming and less precise than with non-virtual environments.
Since global enterprises, almost without exception, run ERP systems, this difficulty with virtual environments contributed to the sense RPA was best suited for specialized and limited roles in automated solutions.
Related Article:Retail Search Technology: Crown Jewels and Dirty Little Secrets
However, while a great deal of attention was being paid to technologies on the RPA horizon, such as: AI, Cognitive, Machine Learning and Digitization; many observers of the industry overlooked the significance of robotic software innovations aimed at expanding the role for robots in business processes and solving the riddle of virtualized environments.
The point of autonomous automation is to maximize robotic efficiencies and work volumes by minimizing the need for human roles. Two things are necessary to achieve these benefits: high precision robots specifically designed to for back office work and a centralized server capable of synchronizing pools of robots into end-to-end processing agents.
Today an autonomous solution is available and capable of running robots and orchestrating work files in a manner that closely resembles batch processing. High precision back office robots minimize exceptions with highly accurate integration and employ self-resolving exception routines to further reduce any need for human interaction.
Central servers are available which not only orchestrate work with robots, but also deploy, monitor and manage robots with protocols that meet enterprise security and compliance requirements. The benefits of autonomous automation are driving global enterprises and service providers to embrace RPA technology, with the Everest Group stating RPA is “the next imperative” for global in-house centers (GIC).
Advanced Image Recognition
This robotic software innovation was the key to unleashing the power of autonomous automation by creating ease of integration with ERP systems. With advanced image recognition, it became as simple, easy and precise for RPA to integrate with virtual environments as any desktop application.
At the heart of this innovation is a powerful image recognition engine, capable of viewing the virtualized ERP snapshots, then finding and capturing the required screen images in a matter of milliseconds.
But what does this mean? Is robotic automation really beyond just talk about transforming the way work is done? If the technology is beyond just talk, is it actually deployed and doing that work? The answer would appear to be yes; according to the Everest Group, 78 percent of GICs have either implemented RPA or are actively planning pilots.
What About AI & Cognitive, etc.?
But what of the exotic technologies so many thought leaders are waiting for: AI, cognitive, machine learning?
The answer is that they’re important and will broaden RPA’s footprint in ways that are hard to fully envision. Ultimately, robotic process automation will be a platform for these technologies, with GICs, third party providers or internal centers of excellence using the best combination to meet the needs of their customers.