Mark Twain had said, “Continuous improvement is better than delayed perfection.” This sets the context for my argument that Intelligent Automation (IA) must precede Artificial Intelligence (AI). IA prepares the ground for sustained benefits with AI, without which AI will not yield the expected returns, or in other words the “perfection” that we are seeking.
Should IA come before AI is a conundrum that faces both our customers and organizations like ours -- how do you make automation intelligent before adding insights and intelligence into the process? Ideally this should come from a combination of wisdom and the experience of dealing with other customers and processes in similar and related domains. The processes that our clients run are unique to their business and therefore they expect solutions that cater to their specific needs. Today there are best practices that deal with specific scenarios and are being built into processes, systems and applications design to help accelerate their AI journey.
An example of this is a leading new online retailer that wanted to set up customer service operations. The company needed to establish a quick service function and associated processes. However, given that it is a new retailer, it did not have any patterns or insights into customer behaviour and required to first acquire a significant amount of customer data. The company needed the automation of its operations and back office to be backed by customer intelligence. It needed to incorporate best practices already available in the industry into its processes and thereby build a continuum on which intelligent automation leads the way to data and analytics. With this approach, the retailer set up operations within three months as against a standard industry timeline of 12-18 months.
A common mistake in AI is over-engineering or automating every process without the foundation of a business case on the transformation. Today we have a plethora of automation technologies and tools but they call for careful evaluation before implementation. Is there a commercial justification for removing the human element entirely? Not all businesses require end-to-end process automation. Businesses which tend to scale quickly will definitely require human interaction and support.
The buzz around robotics process automation (RPA) is drawing the attention of most companies, and rightly so. The difference between RPA and the earlier tool-driven automation is the embedding of intelligence through a combination of machine learning, algorithms, patterns and analytics that together removes process redundancies and waste. There is a strong business case for RPA for organizations that are looking to scale their operations and drive value from their existing operation. However, considering RPA as a magic potion to replace human interactions and save costs is not a sensible business decision. The balance between intelligent humans and intelligent robots is critical to help optimize and drive value across the services organization.