Wednesday, December 3, 2014

AI customer service!? Could that be a thing?


                In today’s world of instantaneous internet communication, it is faster and easier than ever to get a message out to consumers. Consumers have an unprecedented opportunity available to get their opinions and thoughts about a company heard by that company’s employees and leadership. Now that it’s easier on both sides to communicate, the question becomes how can the company keep up? As your company starts to reach a steadily wider audience of consumers soon your ability to respond to messages will be outpaced by the number of consumers sending messages in. Do you just hire more and more people to handle responding to these messages or do you just ignore them? Do you retain a small team of people to respond to the messages but only ever reply by prewritten messages that have no real relevance to the consumers comment? That would save a lot of time so you could keep your team small. With the progression of new technology I think there is a different option.

                Who has not heard of predictive texting? Almost all smart phones available today have some form of predictive text. This is where the smart phone will give you a few options of what it has predicted you are intending to say. It does this by having a database of words and phrases that are often used in conjunction. I just typed into my pre-installed smart phone keyboard the word “how” and one of the options it predicted I might like to use was “are.” I selected “are” from a list of 3 words and it then predicted that I might want to use the word “you” next. “How are you?” Now this is incredibly simple predictions. There is a digital keyboard company called Swiftkey that sells a digital keyboard for your phone that will replace the pre-installed digital keyboard. Swiftkey’s keyboard asks you if it can access your social media posts to learn words and phrases that you often use. I’ve used this keyboard and it would successfully predict some of the most interesting words that are very specific to how I type. It did this seeing that I had used these words and phrases often in other forums and predicted them when I was typing text messages.

                I would sometimes pull a prank on people I associate with by typing a text message that was one hundred percent predictions from the phone. These texts would always be funny because they were so close to making some kind of sense but they were just outside of the coherent range. Swiftkey and Intel worked together to make a technology for Stephen Hawking that will help him write faster by learning how he likes to talk and predicting the words it thinks he would like to use next. As this technology progresses it will be useful in more and more situations.


                Imagine if a technology like this started to predict fully developed responses as it is receiving communication. What if you set your phone to auto-respond to text messages? Maybe someone texts you asking how your day has been, and the auto-respond app looks up what you had scheduled that day and responds, “I’m doing good, but I’m really busy. I have a lot of things going on today and it’s pretty hectic.” As this technology progresses perhaps you could implement it to respond to your consumers in a more personal way.

That’s a funny thought that a message might feel more personal coming from a computer instead of a person.

If you have too many messages sent in from consumers and you can’t respond to them, likely you’re just going to send out a generic message that doesn’t take into account anything the person said. Maybe you’ll send them a few coupons to try and make them feel special. If this Swiftkey and Intel idea gets furthered you could have a computer read the message and respond in a very specific and appropriate way. If you set the computer to wait a few minutes before sending the response it would even feel like the concern or comment the consumer had was important and was being addressed quickly but thoughtfully. It would be infinitely cheaper than paying humans to do the work. Just have a few people monitoring the messages it’s going to send out and you have an inexpensive and effective way to respond to consumers that feels personal. The opportunities for changes in communication through technology are astounding.