Robotic Automation

Robotic process automation (RPA) has been getting a lot of press lately. Some of this attention has more to do with a heightened general interest in "robots" than the actual technology, but most is focused on the compelling benefits RPA can bring to business operations. Acquiring knowledge and a grounded perspective in this burgeoning field is a challenge – particularly (though not exclusively) for newcomers. The objective of this article is to provide a view of robotic software technology grounded in historical context, and framed to support a useful evaluation of upcoming inn ovations.

Becoming conversant with robotic automation technology is a challenge for several reasons. The most apparent one is the rapid pace of innovation. (At UiPath, it’s startling to compare the new features and functionalities in recently released V8.0 with last year’s V7.0, and this pace of change is happening across the industry.)

By moving off the desktop and onto an enterprise platform, RPA benefits have severely undercut the BPO business model: even the best labor arbitrage can’t beat a no-labor arbitrage.

Another challenge is how easily basic RPA capabilities, benefits and requirements can be captured in a simple sentence: Robots mimic human task behavior by following business rules within a modeled process, replacing salaries with license fees to deliver 60%+ savings – and robotic accuracy, performance and scalability is through the roof.

Imagine trying to stuff an ERP or BPMS overview into one sentence.

This simplicity is a challenge, not because it’s inaccurate (it’s not), but because it easily obscures larger and more important truths from the neophyte. These truths are tied to new features and capabilities made possible by significant innovations. As a result, Robotic process automation has gained the power to reshape an entire industry.

This is the proper perspective from which newcomers should view the technology and apply to their own circumstances.

The RPA Capability Curve

When it comes to answering a question often heard from people just coming across this technology, "Why are people talking about business process automation? Businesses have been automated for decades!", the RPA Capability Curve goes a long way.

This observation is true. ERP systems first began lumbering across the corporate landscape in the 90’s. At the turn of the century, business process management systems (BPMS) emerged as a more focused and specific process-level alternative to ERP’s monolithic approach. It’s understandable for people to believe no significant processes were left untouched and un-automated. But many significant processes were left untouched – for good reason: they were ultimately less significant than the investment dollars required for an ERP or BPMS investment.

Instead of automation, companies decided to wring costs out of these processes another way, which became the catalyst for a new services industry: business process outsourcing (BPO). Global delivery and labor arbitrage became the route by which savings could be achieved without investment.

Meanwhile, robotic process automation remained almost unnoticed as it continued to steadily climb its Capability Curve (see chart below).

[click to enlarge]

Desktop

This initial stage remains the dominant image of the technology. A single user, formerly swiveling between screens to manually cut/paste data fields between systems – now freed and focused on other tasks. Or possibly another job entirely. Desktop automation is also referred to as single point automation (see chart below).

[click to enlarge]

Every business process is a collection of activities, and activities are aggregations of tasks. Tasks are work units small enough to be completed by one person but large enough to warrant managerial supervision.

Single point automation is an isolated instance of automation, as in the example of moving data fields between systems. Substantial benefits will be realized, but only in the scope of the task.

Enterprise

This game-changing innovation enables the technology to be moved off the desktop and onto a server-based enterprise platform. The benefits of platform process automation are orders of magnitude beyond desktop, particularly in its ability to assemble groups of robots to perform large volumes of work.

By moving off the desktop and onto an enterprise platform, RPA benefits have severely undercut the BPO business model: even the best labor arbitrage can’t beat a no-labor arbitrage.

The benefits of platform automation are orders of magnitude beyond desktop, particularly in the ability to assemble groups of robots to perform large volumes of work.

Platform automation should be architected for compatibility with customary corporate requirements for security, permissions, compatibility and support. This architecture means the platform can not only provide process automation, but support enterprise shared services or centers of excellence initiatives.

Platform automation should also be architected with capabilities to handle large volumes of work. These capabilities typically include (see also chart):

• Orchestration: in order to handle and complete large volumes of work, the platform must be able to deploy robot groups of varying sizes (from dozens to hundreds) for the purpose of orchestrating and bringing to bear their collective automation capabilities.
• Deployment and Retirement: the platform should have an intuitive management dashboard for quick and easy robot deployment/retirement responses to fluctuations in demand.
• Monitoring and reporting: the management dashboard should have the capability to monitor real time performance of all deployed robots. In addition, a portfolio of standard reports should be available to measure service level outcomes.
• Compliance: a document trail should exist for all robot activities. A common approach is to pull and aggregate log files from all operational robots, then store them in a repository on the server.

[click to enlarge]

Cognitive/Artificial Intelligence

Enhancing robotic software’s rules processing capabilities with cognition will be the next game-changing RPA feature. Robots will work with unstructured data (e.g. voice recognition, email, documents) and be able to independently modify process rules based on outcome experiences or additional knowledge. Many complex and variable processes (think nonstandard claims adjudication) will be within the scope of this technology. Artificial Intelligence is a huge leap and purely speculative at this point in time.

What to Look for in RPA Technology

Anyone with a serious interest in robotic process automation should maintain a steady focus on technology. Keep in mind technology is the only reason attention is being paid to this area. It will be difficult to recover from a poor technology choice. The following points aren’t exhaustive but should prove useful.

• Challenge limitations: competitive software should easily integrate with applications and databases across multiple layers. Specialized interfaces with major 3rd parties (think SAP/Citrix) should be standard. Don’t accept limitations without checking around.
• User experience is critical: it seems every RPA product vendor claims to deliver an intuitive user experience. Don’t pay attention to anyone except yourself and your people. Spend time with demos and draw your own conclusions.
• Enterprise automation a must: don’t settle for a product line without an enterprise platform for automation. Most customers start with a PoC at the desktop automation level, so the product line should allow them to seamlessly move up in the future and keep the value of acquired skills.
• Consistent innovation: your software provider is not far removed from being a de facto partner, so choose one with a track record of consistent innovation. If you don’t see signs of it now, odds are you won’t see signs after you bought the product. Look for evidence in current and prior release notes.
• Good support and happy customers: these two go hand-in-hand and you’ll want a high comfort level in both areas.

[eventpdf]