AI has a long history as a marketing buzzword. But while we’ve been through a number of boom-and-bust hype cycles around artificial intelligence, recent technological developments and the increased availability of systems leveraging artificial intelligence have made it impossible to completely shrug off AI as vaporware or a fad.

It’s (finally) time for brands to start thinking about marketing with AI. So, let’s take a quick and dirty look at artificial intelligence—what it is, why you should care, and what it means for your business now and in the future.

Appboy’s VP of Product Kevin Wang Discusses AI at LTR 2017

1. AI can mean multiple things

When people imagine artificial intelligence, they tend to think of its depiction in pop culture—you know, futuristic robots and malevolent computers, all biding their time until the humans aren’t paying attention so they can take over the world. But while that’s the popular image of AI, when we talk about artificial intelligence in the context of today’s technology, we’re not talking about Terminator-R2D2–Blade Runner AI: we’re talking about something different.

What is AI, really? It’s any artificial system that’s capable of performing tasks that, traditionally, people had to do.

2. Current AI isn’t General AI

Pop culture generally depicts AI that’s the mental equivalent of a human being in all meaningful ways. That’s “General AI.” We don’t have that yet—and if we do get it, most experts think it’s decades away. What we have instead is something called “Narrow AI,” which is powerful but more limited in scope. Narrow AI often means machine learning, which leverages models that automatically adjust themselves based on inputs they receive. Machine learning’s a subset of AI, but an important one, since most of today’s tech uses it.

The key thing about Narrow AI is that it’s fundamentally a system that’s capable of performing very, very specific tasks as well or better than a human being could. And that, we have in abundance. Think Watson, or self-driving cars. Think AlphaGo, which is specialized to play board games. These are powerful tools, but they’re specialized—AlphaGo can’t pilot your self-driving car (and shouldn’t, frankly).

3. AI can’t do everything

The thing is, a narrow AI like AlphaGo is designed to carry out one very specific task and to do it extremely well: win at Go. There are rules that govern the game, with strategies and possible situations that govern different outcomes, and AlphaGo’s on top of all of that. But, ultimately, it’s a one-trick pony—even if it is really, ridiculously good at that one trick. Try as it might, your self-driving car is unlikely to beat you at Jeopardy, at least in the near future. And that’s true of basically all of today’s AI.

4. With AI, it’s all about the goal

IBM’s Deep Blue (Source: The Verge)

When it created AlphaGo, Google’s goal was to create a system that was incredibly good at playing Go—and they succeeded. Same thing with Deep Blue and chess, or personal assistants like X.ai and scheduling meetings.

But while Narrow AI is great at reaching those kinds of focused goals, it’s not good at picking goals for itself—which means that the success or failure of modern AI depends in part on how good a job the humans overseeing the AI did at focusing it on the right task and feeding it the right data.

5. You’re already using AI

Gmail’s AI-powered email spam filter

That’s the thing about AI: it’s exciting as long as it’s new and innovative, but as soon as it’s not, people start to go, “Wait, this can’t be AI—this is everywhere.” And it really is everywhere. Email spam filters. That’s AI. Amazon’s product recommendation engine? That’s AI. Even smart-dragging a series of numbers in Excel is technically a simple rule-based AI. Face it—AI’s already here, and you’re already reaping its benefits.

That’s especially true if your brand takes advantage of Appboy’s lifecycle engagement platform: when you use Intelligent Delivery to make sure that your messages reach each customer when they’re most likely to engage or leverage Canvas with Intelligence (which lets you orchestrate smarter customer communication), AI’s making that possible.

In general, though, we only call something AI if it impresses us. No one says, “Wow! Spam filters! I’ve seen the future and it’s incredible!”

6. Mobile is driving an explosion of AI development

You know what AI loves? Data. It’s basically the fuel that all of today’s artificial intelligence systems run on. Without significant high-quality data, it’s hard—or impossible, even—for AI to accomplish anything significant.

One big factor in the recent vogue for AI is the massive increase in available data that came with the rise of mobile and a more connected world. It’s no big secret: customer data collection is easier—and more detailed—than it’s ever been. People carry their smartphones with them everywhere, even sleep with them, and if you’re a brand with your location data game in order, it’s possible to target customers for messages based on where users live, where they go, how often they engage with your app or website, and the products and services they care about.

All of those data points can be fed into a well-designed AI system, and with those kind of massive data sets at their disposal, it’s possible for an AI to identify patterns or pull out insights that a human marketer would miss—and to do it almost instantly. Those insights can do a lot to make your marketing efforts easier, smarter, more effective.

7. AI isn’t a magic bullet

I know we’ve covered this, but it bears repeating: AI isn’t like in the movies.

You can’t just go to your computer and say, “Okay, Google, make my marketing better,” or “Alexa, sure would be great to have higher user retention,” and have it happen. You have to prepare your AI with high quality data, and train it to accomplish the specific, task that you care about before it’s going to be able to do anything impressive.

Maybe that’s a little disappointing to hear, but let’s be honest: even if your company had access to the kinds of artificial intelligence you see in the Terminator movies, would you really want to leave your customer engagement strategy in Ahhnold’s hands? (Don’t think so.) That’s the temptation and the danger of AI for a lot of companies—there’s something seductive about imagining that you can just sit back and let computers make your marketing work. But that’s not real life. And if it were real life, you’d probably be out of a job.

8. Before taking advantage of AI, you need to nail the fundamentals

AI can be really cool. Intelligent Delivery is a really powerful tool, and Appboy’s new AI-powered Canvas features are going to make a lot of marketing campaigns way more effective. But here’s the thing: to fully leverage AI, you need to do the basic things right.

Using AI to optimize your brand’s promotional outreach can do a lot to boost your ROI, but not if the people who download your app don’t stick around long enough to receive any of those promotional campaigns. It’s not rocket science. You’ve got to walk before you run, and you’ve got to get the basics right before you can make the most of artificial intelligence as part of your customer engagement efforts.

9. To make the most of AI, you have to identify places where humans and computers can work together

Today’s AI is great at processing vast quantities of data to find hidden patterns and actionable insights, and making high-speed decisions in connection with a specific goal. People, on the other hand, are way better than computers at creative thinking. And that’s a great thing.

AI shouldn’t replace your marketing team—it should make your marketing team that much more effective. That’s the thing about a great AI system: it’s like having a really amazing hunting dog. Lassie knows how to smell where the foxes are and how to track them down, but she’s got no idea what to do when she finds them, or why, exactly, we’re hanging out in the woods in the first place. That’s where people come in.

AI is the muscle. We’re the brains of the operation.

10. Right now, AI is a nice-to-have, but the time will come when it’s a necessity

It’s incredible how much data marketers have at their disposal today, compared to ten, twenty years ago. And the tools we have for reaching customers would have sounded like something out of science fiction years ago. But, the way things are going, in another ten, twenty years, we’re going to look back at 2017 and think “how did marketers ever do their jobs without having the technologies I’ve got now?”

As time passes, technology always makes things faster, easier, and more efficient. And while we grapple with the challenges and opportunities that are going to come with tomorrow’s larger, even more detailed customer needs, AI is only going to become more essential to the way companies interact with their customers. Right now, it’s totally possible to provide your customers with a best-in-class brand experience without using any AI. But in 10 years, who knows? The companies that pay attention now and start thinking about how they can use AI to build stronger, more sustainable customer relationships today are going to have a major advantage in tomorrow’s marketing landscape.

You know, if the killer robots don’t get us first.