Who Knew Clippy Was So Smart

I love automation in product service delivery, especially when deployed in consumer workflows that should be frictionless at this point. Understanding where I ship, how I pay, my sizes, what components fit my car or computer as I shop, or assisting me with lower level support issues using natural language processing are some examples where automation and bots should rule the world today.

But, from a user experience point of view, many of the AI bots and virtual assistants I am encountering in actual consumer products today (not the cool YouTube demos everyone posts) feel like a step backwards. They even make me miss Clippy, who was vastly more useful compared to some of today’s automated tools. Yeah, I said it. 

Bots “powered by machine learning” can be a minor annoyance in the consumer goods space with experiences such as repetitive logic loops when unexpected inputs are received, limited capabilities to understand individual preferences, and forced segmentation that does not fit my user story (you will learn about Product X no matter what you asked Ms. Bot about Product Y). They are also far from ready for prime time when it comes to natural language responses, unless you like to be forced into reducing your vocabulary into a series of two or three word queries. Again, this is from the point of view of the actual user who is interacting with every day services (shopping, content searches, virtual assistants, home automation, chatbots, help desks) not the futurist who is focused on over the horizon capabilities.

These same minor issues in the consumer space could be catastrophic when brought to the health care sector. A poorly deployed chatbot could lower already dreadful engagement and lead to increased friction in care—the opposite of the goals outlined in redeveloping health care delivery systems for the future.

Humans trust humans with their health care, and that is ok. In fact, for our behaviors to modify and change over the long term, other humans are still key. So when applied to health care, bots should work to automate low level workflows and scheduling with the purpose of increasing the value and seamlessness of human to human contact. They should optimize the time spent with your care givers via text, phone, email, video, or in person by getting you to the right one, at the right time, based on your preferences and needs.

Serious development around automation with bots in health care is exciting, but they need to be applied in a manner that is far more empathetic, individualized, and purposeful than what we see in the consumer space today. And they need to be deployed with the realization that behavior change in health care requires human support and interactions-even Clippy would have guessed that one.

Related:

• Six year olds smarter than current AI solutions: https://www.cnbc.com/2017/10/02/google-ai-has-almost-twice-the-iq-of-siri-says-study.html