The Growing Symbiosis Between RPA and AI
Evidence suggests systems in which the two automation-focused methods work together and become increasingly common in the not-too-distant future.
Join the DZone community and get the full member experience.Join For Free
Given their inherent similarities, it’s entirely logical that robotic process automation and Artificial Intelligence would crossover in a variety of different contexts. Both of these technologies are contingent upon bringing order to processes — usually workflows, in either the physical or digital realm — that might otherwise run the risk of falling into disorder for any number of reasons, most notably, human error. RPA and AI are perfectly capable of operating on their own, and in fact, often do. Yet recent evidence suggests systems in which the two automation-focused methods work together and may become increasingly common in the not-too-distant future.
RPA and AI are working alongside each other in various business applications.
Organizations considering implementing one or the other to manage any of their operations would serve themselves well by examining how they’re being used in tandem. It’s also worth investigating how a versatile low-code business application development platform like Appian may be ideal for creating tools to most successfully leverage symbiotic RPA and AI.
The Crossing Paths of RPA and AI
According to InformationAge, it can often be easier for most organizations to implement RPA in advance of AI, especially in instances where they are still reliant to some degree on legacy systems: hardware, software or both. Complete overhauls of companies’ existing tools for business process management in favor of more advanced technologies are quite difficult, as they can’t be accomplished quickly — and, in some instances, aren’t realistic at all. The better approach, then, is to use the addition of something like RPA as a table-setter for more substantive changes to operating methods.
With that being said, the increasing frequency with which RPA and AI are coming as a package deal means BPM advancements or overhauls won’t be quite as complex to manage in the next few years. AI-enhanced RPA platforms that automate various organizational workflows and processes either in whole or in part, referred to as CPRA — cognitive robotic process automation — are gradually becoming the norm, in yet another example of how truly pervasive digital transformation has become across multiple segments of both the public and private sectors.
RPA-AI Hybrid Solutions in the Pharmaceutical Industry
Without a doubt, the business of pharmaceutical development and production is one of the most highly regulated sectors in the world, in no small part because it’s characterized by numerous complex processes. One mistake in the chemical formulation of a particular drug — or a flaw in the physical steps of manufacturing tablets, capsules, liquids or other forms of medication — can lead to illnesses and even deaths among users, which can open up the drug company that is responsible to a surfeit of civil or even criminal penalties, in addition to various regulatory sanctions. Because of this, it’s of the utmost importance for pharmaceutical businesses to implement numerous fail-safes in their BPM practices, and the combination of RPA and AI can help provide these precautionary measures.
According to Robotics and Automation News, major companies in this field may well consider this path in light of the success recently seen by AstraZeneca, one of the world’s biggest pharmaceutical producers, when it commissioned the services of Deloitte in pursuit of an RPA implementation. While this would by no means be the first time that an organization within the general sphere of health care adopted RPA, CRPA, AI or some combination thereof, no one in the pharmaceutical field had tried using such technologies to handle the specific task of “adverse event reports.” Other pharma businesses might use a different term for these filings, but the meaning is likely the same: reports of detrimental side effects in patients who have used a particular drug.
Cognitive RPA improved the adverse events reporting processes of a major pharma company.
AstraZeneca receives an average of 100,000 adverse event reports each year — many documenting minor ailments but some detailing serious illnesses. Before the firm enlisted a major business solutions provider to devise an advanced RPA system for this purpose, it dealt with these reports manually. As such, members of the AstraZeneca patient safety team were spending millions of hours every year personally administering whatever tasks were necessary to ensure patients’ adverse experiences were dealt with appropriately. Because individuals’ health is at stake in such cases, care and sensitivity are essential – not to mention the initiation of any adjustments that can be made to the drug – but certain routine administrative processes could be (and were) automated through the RPA platform.
Robotics and Automation News noted that this RPA use case was a considerable success, bolstering response rates from both health care professionals involved in drug trials and AstraZeneca staff. Although similar solutions will have to undergo rigorous compliance testing via computer systems validation, as this one did, high-level CRPA application platforms will likely be adopted more broadly throughout the sector.
Driving Advances in Voice-Activated Assistants
According to a January 2018 study conducted by Capgemini, 24 percent of surveyed individuals stated that they would, given the opportunity, use voice-activated digital assistants or chatbots to complete various purchases rather than doing so through direct interaction with a website. Additionally, the research found that Amazon Alexa, Google Assistant, Apple’s Siri and IBM’s Watson, along with any other similar applications that will likely emerge in the next three years, will stake claims as a dominant channel of commerce for average consumers. Even if voice-activated assistants don’t quite reach heights that can be considered “dominance,” it’s practically a given that they’ll grow in popularity between now and 2021 or 2022, as their price points go down and they become accessible to a wider range of customers.
The symbiotic operation of RPA and AI grants these systems their effectiveness. An AI platform, as present in any of the aforementioned assistants, can generally understand a typical human question or command — “Siri, set a timer for two hours,” or “Alexa, what’s the current record of the Boston Celtics?” — but an RPA cannot. Conversely, the AI can’t complete the nuts-and-bolts processes handled by the RPA. When combined, however, the AI translates the voice command into a series of simpler signals, sends them to the RPA as requests to locate relevant data packets, receives this data and finally converts it into a natural response: “Timer set for two hours,” “The Boston Celtics are 9-8, 6th place in the NBA Eastern Conference” and so on.
As RPA and AI develop further in the years to come, their capabilities will grow: AIs will potentially catch on to the slang and syntax of their users instead of requiring that questions be asked in formal sentences, while RPAs grow capable of collecting more complex requests based on prompts. As a result, their fusion can only bring about greater advances, in voice assistant technology and much, much more.
Benefiting Customers and Workers Alike
It’s not uncommon for employees to fear that automation will marginalize or eliminate their jobs. However, as Forbes Technology Council contributor Kris Fitzgerald points out, this may not be the case in most industries: Hybrid RPA and AI systems take the burden of paperwork, data entry and other time-consuming tasks away from employees — particularly those dealing with customer service, in numerous sectors — and therefore allow them to put their intrapersonal skills to greater use. Fitzgerald noted that hotels have seen a great deal of success employing RPA and AI together for various customer-facing needs, and that managing invoice processing represents another notably broad use case for these united technologies.
Opinions expressed by DZone contributors are their own.