About dating, they say there are a lot of fish in the sea. This is a handy metaphor too for thinking about apps, those little software packets swimming just under the surface of our interface. There’s one that will call a person to pick me up and take me places. Another one allows me to order meals from a variety of restaurants and grocery stores. Still another finds people for me to try to sleep with. Used to be you’d have to fish each one out of the system. If you wanted to use a few at a time, you’d have to pull up each one individually.

But suppose you had a sort of virtual fisherman there with you, ready for you to tell what kind of app you wanted pulled up from the deep. It would connect the functionality of all your apps, so that when you wanted something from one of them, you’d wouldn’t go to the individual app, you’d just ask your personal fisherman. That’s a bot.

Bots are, simply, software used for executing a series of steps. But now we see them streamlining tasks you’d have to do yourself: ordering a pizza, scheduling an event. They’re taking over the responsibility of fishing each app out of the depths. Once they get them above the surface, they can also connect the apps, facilitating interaction while we kick back.

A Tool for Efficient App Wrangling

In order to understand how bots and apps will work together, I talked to Chalenge Masekera. He’s a data scientist at Salesforce. He builds bots, and the impression I get is that he, like a lot of data scientists, is very busy (there’s a shortage; the industry needs 190,000 more). I begin to worry that he is so busy that he might not have time to talk to me, and will build a chatbot to answer my questions — a Chalengebot that would tell me about analytics and metrics. To prevent this, I make sure to get him on the phone. (It would’ve been interesting, admittedly, if I had built a bot to ask questions about bots and he’d built a bot to answer questions about bots and then we could’ve sicced them on each other and gone out for a sandwich. But the world isn’t ready for that yet.)

“When you talk about an app, you’re usually talking about a mobile app or a web app,” he says. “So there’s a user interface in which you perform a series of steps. Let’s say you want to summon a cab. You open up your app and you manually go through a series of steps. What bots essentially do is provide instruction for all those steps that you’re performing, automating that series.”

In that sense, the bot is the middleman between you and all your apps. “What’s happening with bots is easy and seamless bundling of functionality across multiple apps,” Masekera says. “You don’t have to switch from one app to the other. You can — with one bot — order your pizza, order your Uber, check the weather, check your account balance, check the top news headlines from there. Almost every functionality is available.”

“You don’t care whether you’re actually using an interface or you’re texting. Now, you’ll no longer need to use that app. All you need to do is interact with the bot.” 

— Chalenge Masekera, Data Scientist

I imagine how a bot would work for me. I could ask it to find me a date, work out the details, put it on my calendar, and then afterward it could find the nearest place to order sad take-out noodles for one, which I would sadly eat in the backseat of the car that the bot had called for me. Probably this bot, built with rudimentary AI, would be able to look at my record of success on the dating app and predict how the night was going to go (This date will not extend past one hour, the bot will think). The bot’s AI would have the noodles and the car on speed-dial, so to speak. The bot could be called something catchy, like “L0nr,” where the “o” is a zero, to really emphasize.

There are, of course, bots that dispense with the need to date humans entirely. Bots come often in the form of chatbots, messaging-based programs that facilitate customer service (like that pop-up that fields questions on a company’s website), or planning your days and nights. They seem sort of like people, but people who can be easily confused. They haven’t quite passed the Turing Test yet. One can see how bots will be closer to us, like friends and assistants, helping us handle our affairs. That could be the personal (ushering us through a mostly unsuccessful love life) or the professional. As bots get smarter, they’ll take on more responsibility.

Better Friends

“They can go from having very mundane tasks to more sophisticated ones,” says Masekera. “For a salesperson, it could be something that performs simple sorting or filtering operations. But it could also use machine learning on your leads. You could say to your phone, ‘Show me leads I’m supposed to talk to today,’ and it does those operations for you: analyzes which ones are at which stage, finds the hot leads, and gives you a probability of which ones you should talk to first.”

“Or,” he continues, “let’s say you are a marketer and you send out specific emails at specific times, and you don’t want to go through the same repetitive steps over and over again. A bot could do that for you.”

The bot, then, extracts the valuable part of the app (or apps) by taking care of the manual part of the interface.

“What you need from apps is to be able to perform a series of tasks. You don’t care whether you’re actually using an interface or you’re texting. Now, you’ll no longer need to use that app. All you need to do is interact with the bot,” he says. “Bots might replace a lot of apps, but I think there are still a lot of apps which I don’t think bots are well-suited to performing, like Instagram.”

Masekera reminds me that when apps came around, they were heralded as the death of the desktop. But that hasn’t happened just yet either.

I like to think that my newly designed L0nr bot will learn to be a better friend to me in my solitude. It won’t just order the pizza; it’ll see that I’m going to be alone for many nights to come, and learn that my orders should be the “personal” size.

“It knows what your favorite order is, depending on usage and how many pizzas you normally order,” says Masekera. “We are moving from dumb bots that remember only the last command you sent it to bots that remember how you interacted with them, and are able to learn from that and make more smarter predictions and smarter interactions.”