It’s hard to find a product today that doesn’t claim to use machine learning to provide artificial intelligence (AI). The funny thing is that the marketing has changed while most of the products have not. It’s amazing how so many products have supposedly morphed into AI-based solutions overnight, despite little evidence of any product development effort. This, of course, means that the vast majority of these solutions do not offer AI, and what we’re reading in their marketing materials and websites is merely aspirational messaging.
To be fair, it’s going to be really hard for a vendor to attract attention today if they don’t claim to use AI to improve the performance and output of their solution. Given the choice between purchasing a product that continues to do what it always did or one that uses AI (which typically means machine learning) to enhance its capabilities, most people are going to go for what they believe to be the newest and greatest cutting-edge offerings.
The problem is that most of what people are buying is hype, and this is going to result in disappointment as enterprises realize that AI is really in the earliest stages of commercialization. The potential is great, but the current generation of technology and the applications are far from fully AI-enabled. As I look through the market, the vendors who are closest to delivering AI-enabled applications are selling a great deal of professional services with each solution in order to build out their products.
The driver behind the AI revolution is the need for productivity and quality improvements, which are important for all enterprise applications and essential for people-intensive front- and back-office service organizations. Imagine a voice self-service solution (also known as an interactive voice response system, IVR) that self-corrects when it realizes that customers are dropping out at a certain point in the script (application). If machine learning were applied, the solution would identify the issue by itself and then make a change to the appropriate components of the script without human intervention. Another great use of AI would be to embed it into an automatic call distributor (ACD) to continuously improve and optimize routing. Imagine an ACD that continuously enhances its routing algorithms, ensuring the right transactions are delivered to the best-suited agents/associates. These examples sound great, but are not fully-baked today. Most of what is currently referred to as AI are business rules created and modified by humans. These approaches are not new, although there are changes in how they are being applied and rolled out as vendors strive to make their solutions more intelligent and AI-ready.
Read all at the source