Just as industrial robots sparked a new era of high productivity in blue collar industries, Robotic Process Automation has, over the last twenty years, revolutionized those of the white collar. Now promises of RPA are plateauing, and enterprises are already looking beyond it.
RPA gave business a new way to think about, well, pretty much everything. Whether applied by workers to administer business processes, IT, workflow, remote infrastructure, or back-office work, the two core benefits are the same: First, improvements in accuracy, as human error is minimized. Second, reductions in the amount of time it takes to get work done.
RPA also elevates the nature of work by removing people from dull, repetitive tasks. But that is exactly its problem — the only things it is capable of automating are brainless. Data entry, information capture, button pushing. These tasks involve no intelligence and can be easily programmed. The stuff RPA does is, literally, robotic.
All these benefits aside, RPA’s capabilities are limited to tasks that involve no intelligence and can be easily programmed. The question arises whether it’s possible to achieve the same kind of productivity gains in intelligent tasks?
Enter Cognitive Technology
The answer is a resounding yes. Cognitive Automation brings in a whole new level of sophistication.
While most workers now are are freed from the most mind-numbing of tasks, they are still chained to workflows far below their intelligence. A procurement manager no longer has to transfer service provider pitches into a central document, but she still must waste an entire day looking through the list an RPA tool has given her in order to narrow it down to a shortlist of best options.
There are an abundance of value-added tasks she could be doing — meeting providers in person, building industry know-how in her internal team, aligning services to long-term strategy — which would give her employer higher ROI. But she can’t. She is still in a situation where precious time and resources are consumed doing things manually, choosing from limited options and making sub-optimal decisions.
To enable humans to engage only in work that corresponds to their full intelligence, we need to be able to call on machines much smarter than those that fall under the RPA umbrella. Machines that emulate more closely and enhance human thought process. Enabled by advances in the technology of cognitive computing, Cognitive Automation is driven by artificial intelligence that can build knowledge, understand natural language, interact naturally with human beings, and overcome ambiguity and adapt to new types of problems. It’s configurable AI for human language.
Cognitive Automation excels in tasks involving masses of unstructured data sets. These tasks typically require decision-making that can be self-taught and a human-like understanding of context. Our procurement manager for example would not only be given a shortlist of best options based on pitch data — she would also get, thanks to Cognitive Automation, an intelligent feed of external data like social media feedback from existing clients or analytical coverage in trade magazines that give a more rounded view of the quality of a provider’s service.
In the financial sector, we can imagine Cognitive Automation as a team of interns, 1,000s-strong, crunching through a million documents each day, finding the five-to-ten bullet points that will be relevant to each of a stock market investor’s decisions. In retail, as an intelligent assistant intuitively stepping-in to influence consumer buying decisions. In healthcare, as the most well-read doctor in the world, annotating pathways at life-saving speed. Cognitive computing is the key to unlocking the automation of complex workflow.
Example — Document Redaction
To explore a use case that can be applied to almost every enterprise, especially those in the finance, tax, legal and government sectors, let’s look at document redaction.
Being equal parts important and laborious, redacting sensitive information is a universal pain-in-the-ass. Not only is it expensive and a major bottleneck for knowledge-sharing within and outside of organizations, it’s dangerous — human lapses in redaction account for 25% of all data breaches (40% in the case of governments). Global cybersecurity risks are growing increasingly severe — yet, in the case of redaction, most of these breaches are inadvertent or easily avoidable.
RPA can already:
- Redact anything that follows a certain pattern, like a social security or credit card number.
- Redact anything with a repeating pattern, like a name.
- Redact all names given in a list; clients, potential vendors, mergers and acquisition targets, and so on.
Such software improves security and saves a lot of time. A business can be 100% sure a document it shares with a marketing affiliate contains no social security number, while the finding and redacting of that number is completed in a flash.
However, redaction is not always that straightforward. Decisions need to be made based on the context. For example, in a sentence like “President lives in the White House”, there is hardly anything that needs redaction. However, “The president met with Mr.Giuliani at the White House.” may call for redaction of “President”, “Mr. Giuliani” and “White House”. Or, consider the token “35%” in the following sentences — “IRS’s maximum tax slab is north of 39%”, and “Apple’s offshore cash reserves are 35% of total assets.”
Building upon RPA
Cognitive Automation builds on RPA’s qualities and introduces an extra level of sophistication; contextual adaptation. Like a business adapting its strategy to dynamic market conditions, Cognitive Automation can adapt the rules it uses to redact information depending on the evolution in the context of the data and workflow it processes.
When combined with other cognitive capabilities, this creates a much more intuitive form of enterprise and an information architecture humans can easily and naturally interact with. For example, a chatbot could sit between the worker and the information they need: The worker requests the information using natural language search and the chatbot, recognizing the worker’s relevant position and role within the organization, retrieves the documents which contain that information, automatically redacting any other confidential information that happens to be included alongside. If someone more senior looks at the same document, they will be able to see more of the information.
Cognitive Automation is not without its downsides. It is a nascent industry, and programs can take anywhere from four weeks to two years to configure and train.
But, if business is willing to look beyond the hyperbole and think about how this technology can be applied tactically within the organization, it will reap the rewards successful application brings; the enabling of workflows that were once prohibitively expensive, took too much time, or were overly-dependent on the people completing them; and the freeing of human capital to focus on more complex and value-added work.
In short, cognitive automation is here to complete the revolution robotic process automation started.