In 2012, an article in Wired magazine titled, “Let the Robot Drive: The Autonomous Car of the Future is Here” foreshadowed the coming self-driving car revolution. Now, only six years later, Autonomous cars have a foothold on our roads and each major auto manufacturer has incorporated safety features based on self-driving car technology. Some companies, like Tesla, (see Autopilot avoid a semi-truck) have made self-driving capability a key feature of their vehicles.
Why have we been able to take something as life-critical as driving and actually make a system that improves upon human performance in a mere 6 years?
The obvious answer is that advances in Artificial Intelligence (AI) deserve the credit. While it is certainly true, it doesn’t tell the entire story. The AI algorithms used by self-driving cars are generated via a curated set of training data that comes from sensor readings from actual car trips and events. Furthermore, as the car operates, it gathers the full context of its environment at a speed and distance necessary to react appropriately.
Situational awareness provided by a wide-array of sensors combined with the sophistication of the AI is what makes the Autonomous Car a reality.
So, what learnings can the support industry gather from the autonomous car? Today, most people would claim that some forms of Autonomous Support are already in place in the support industry. After all, chatbots and voicebots are interacting with customers, asking questions, and running down programmed decision trees and resolving issues. But is this really autonomous? As in the self-driving car example, the issue is one of situational awareness. The job of a chatbot is to gather situational awareness and then use that data to do rigid things. The AI part of those systems is focused on language parsing – the rest is programmed based on human answers. It can be useful, but it is not Autonomous and depends on the work of a human (the customer) to provide the sensing.
So how do we make support Autonomous?
I have spent several years thinking about this topic and would propose that Autonomous Support is built on the three pillars:
- Situational awareness gathered via automated means
- Real-time determination of a specific, singular action, using AI based and the gathered data.
- Completing the action to the level that the human counterpart would.
If my definition were in operation, it would be a revelation to the support industry. Today, a lot of work and effort are spent just gathering situational awareness. Chatbots do it, IVR systems do it, everything does it. It is time consuming, annoying to the customer, and immediately creates an adversarial relationship between a person that needs help and the company that can provide it. The lack of situational awareness cripples AI’s ability to find issues and resolve them.
In the coming weeks, I’m going to explore in detail each aspect that together yield an Autonomous Support solution along with the tangible business impacts.