Why robotic process automation is not artificial intelligence
The inconvenient truth is, artificial intelligence has become an overused buzzword. As more and more financial companies overhaul their business intelligence tools and embrace predictive analytics, the term AI has been thrown around so haphazardly that some confuse it with Robotic Process Automation (RPA). The reality, however, is that these two tools are lightyears apart in terms of capabilities and applications.
Robotic process automation – What it can do (and what it can’t)
RPA is defined as the process of using a robot (usually software) to automate repetitive tasks based on a strict set of rules. The keywords in this definition are ‘a strict set of rules’. As CIO Magazine aptly puts it, “With RPA, businesses can automate mundane rules-based business processes”. According to Ernst and Young, examples in the financial services industry include:
- Updating details of new customers
- Financial ratio analysis
- Calculating exchange gains or losses in financial statements
- Generating reports in multiple formats
- Preparing reconciliation statements
Therein lies the key difference. For RPA to be effective it needs a strict set of rules to operate, typically in the form of “If-Then” statements. It’s really no different from the IF function on Microsoft Excel which returns one value for a TRUE result, and another for a FALSE result. So, while offering some advantages, RPA solutions are essentially ‘mindless’ robots capable only of routine, low-level tasks. In contrast, AI is not limited by rules and is capable of making judgments based on analysis and extrapolation of past data. It is designed to augment and replicate intuitive human decision making – this is the AI advantage.
Robotic process automation vs. artificial intelligence example case: Medical claims
When an insurance company processes medical claims, RPA initiatives can indeed save time on clerical work. As there is a clear predefined ruleset, it could route the claim to a specific department, or flag if the medical claims exceed a certain amount, or sort claims by other predefined criteria e.g. type of injury or illness.
Artificial intelligence on the other hand, would be able to analyse the entire claim and then assess whether it should be approved. But its capabilities don’t stop at the beginning of the claims process. Even after the initial claims assessment, AI is capable of detecting subsequent fraudulent claims down the line – giving it the ability to conduct full end-to-end risk management.
The crucial difference here is that this initial assessment and end-to-end risk management performed here is not based on any predetermined criteria. Instead, by feeding the machine past claims data, it can learn from thousands of different data points, make sense of anomalous patterns that emerge, and thus separate legitimate claims from unqualified and potentially fraudulent claims.
AI technology therefore performs the function of a decision-maker, as opposed to a blind rule follower.
A glimpse of the future – Will AI replace robotic process automation?
Because the confusion between AI and RPA still exists, some RPA companies and services are falsely recognised as artificial intelligence. Confusion also ensues because as AI continues to evolve and grow in understanding, many RPA companies are transitioning their business models toward AI. This is a glimpse of a shifting future—one where AI predominates, and RPA slowly fades into the background.
However, ‘fading into the background’ doesn’t mean the total disappearance of RPA. The truth is, not all work needs human-level decision making. There are still many tasks within businesses and the economy that are, at their core, essentially mindless. In those instances, using AI is unnecessary, not because it lacks capability, but because it is overkill. To use HR jargon, AI is simply overqualified for those jobs.
Thus, from a pure cost perspective, companies will likely still opt to use RPA for these low-level tasks. They will use AI for tasks that sit higher on the value chain – those that require human-level judgment. The evolving relationship between AI and RPA could end up looking very much like the relationship between humans and RPA prior to the advent of AI in business. AI handles the complex high-value tasks and RPA automates the low-value ones.
Since AI is able to replicate human decision-making, it is only natural that it will be far more prominent in our minds. Yet, RPA technology is still likely to be there, quietly humming away in the background, doing the menial and tedious admin-type tasks that still need to be done in the modern economy. We’re just unlikely to really notice it.
The future is AI
AI is already a hot topic in today’s world, and for good reason. As the world becomes more complex and AI’s capabilities continue to evolve, its importance will only increase. Forward-looking companies, especially in the financial industry, have already realised this and are aggressively investing in their AI capabilities. The rest are likely to be left behind.
At AiDA Technologies, we are powering leading financial institutions and insurance companies across Asia with a suite of AI tools that accelerate lending and claims processes, identify revenue potential, drive cost reductions, and anticipate evolving risk. With our multi-award winning proprietary machine learning technology our clients are able to unearth unknowns and stride confidently into the future. Contact us to find out how we can help you do the same.