Rise of the chatbots, and why you should care

There’s a little-known app called Mezi that can help you order flowers. You start a chat, explain what you want, and Amazon delivers the order a few days later.

It’s not rocket science. In fact, while there’s A.I. involved in parsing out your order and finding the best options, most of the shopping advice is provided by actual humans.

What impresses me is that it actually works. It saves time. The flowers arrive.

Here’s another example. The other day, I was thinking about getting the new Bill Bryson book. For some reason, I was drawing a blank on the title. I know it’s about England, which is timely, considering the Brexit vote and all of the uproar.

At this point in the grand spectrum of technological advancement, I have a few choices. One is to search Amazon on my own. You need an app for that, or a browser, and you have to type a search phrase like “bryson latest book” and then hope the results make sense. The time-consuming part is when you have to log into a shopping site, find the book title (or flowers to order, or flight to book), arrange the shipping, confirm your credit card that may or may not have expired last week, and click the buy button.

I decided to set a stopwatch for placing a real order on Mezi and one on my own. Since the chatbot A.I. only works for flowers, flights, and hotels, I searched for a flight to San Francisco. My own search took about 15 minutes, start to finish. With Mezi? I asked about a flight from MSP to SFO, and a chatbot (plus a human rep) got to work. While it took half an hour to get the results, it took me 30 seconds.

It’s that part about not really knowing what you want that plays such an important role in the process, and why employing chatbots is so helpful. Facebook is doing it, American Express is doing it, Mezi is doing it. Maybe you’re doing it. A.I. can handle that initial parse, that moment of exploration. It’s incredibly important, though, to understand why this trend is happening — and if it even is a trend — especially when you think about the problems 3D printing is having right now and the hype surrounding virtual reality headsets.

For starters, chatbots offer the best way to deal with overload. They are important because they are paving the way to a more A.I.-assisted future, one where bots fill in the gap between what we don’t know (and maybe what we don’t know that we don’t know) and what we actually know. It’s like the Johari Window (a technique used to help you learn more about yourself) comes with an app. We don’t fully know how A.I. can help.

Think about it this way. Since I’m a major car enthusiast, it helps to understand how A.I. has progressed when there’s a physical product, and it helps when that physical product is a Tesla. Similar to how a chatbot helps us find things, order things, learn things, parse things, and even avoid making mistakes (e.g., Sorry honey, I’ll remember you like daisies next time), a Tesla Model S with the Autopilot mode enabled knows there is a bus coming up on your right-hand side. You’re too busy listening to Radiohead. The A.I. can parse information you don’t even know exists, and in the future you won’t even care that you don’t know it exists.

Back to the Mezi app — it doesn’t quite do this yet, but the shopping assistant could be programmed to remember that my wife likes daisies. It can also store my credit card number. Maybe it will even tie into my bank account someday and update my card number.

The “trend” part? This can only get better, and it will. It’s a bit like speech technology from about 20 years ago. When it comes to speech, the dictionaries have improved every year. The tech gets better the more we speak, and here we are using Siri and Alexa as a result. Chatbots will someday be able to remind me about appointments, let me know that there’s a good deal on the pizza I always order, prompt me when there’s a new Bryson book, and even banter away about Brexit, possibly using correct British grammar.

I like that chatbots tap into a flowing, vibrant, and active nerve — e.g., the total domination of apps like WeChat and Messenger, WhatsApp and Snapchat, and all of the other texting and photo apps that are now becoming unicorns, or already are. The best trends add a new layer to an established foundation. With 3D printing and VR, a layer — such as rapid prototyping and augmented reality — existed (though it wasn’t widely known) to provide a fertile ground. Chatbots can springboard nicely from the texting apps, and they already are, but they have a ways to go.

I want to type questions like: “Keep track of the best flights to Austria over the next four months.” That doesn’t work at all right now, but I’m hoping it will soon, especially since I’m actually taking a trip to Austria in the next few months. Another one: “Use my data from the Fitbit Aria scale to keep a record of my weight loss, and share it daily on Facebook.” That’s a bit complicated but not impossible. A chatbot would have a few questions for me, like which friends should see the posts. And there’s a question about whether that would be handled by a Facebook chatbot or one made by Fitbit. It will be exciting to see how this pans out and what the most creative minds come up with.

That’s why I’m following it so closely. Stay tuned for more.