AI is transforming almost every industry—making things that we’d never even thought possible seem almost commonplace. Whether you’re a marketer using AI to automate tedious tasks, a doctor using it to help diagnose diseases, or a teacher using machines to help evaluate a child’s reading, you’re probably putting machine learning to use in one way or another. Publishers are no exception.

At this point, publishers are used to the industry being disrupted. The internet threw everyone into a tizzy. Then came social media, the mobile web, apps, and now voice search. AI and machine learning have the potential to make publishers more efficient and more successful—and even solve some of the problems caused by past disruptions—but only if you have the resources to invest.

Machine learning can help publishers tag videos and images to make them easier to find and help translate content for international audiences. It can help you experiment with new formats and can even help publishers identify new subscription opportunities. Of course, you need to be able to devote the resources—both financially and in terms of manpower—to make the most of the opportunities available through machine learning.

Media organizations with deep pockets have always had an advantage over mom-and-pop shops, but as changes in technology become more and more rapid—and the need to keep up becomes more high stakes—the gap between the technology haves and have-nots is widening. And the further you fall behind, the harder it is to recover.

It’s an ugly little cycle. The bottom falls out from under the publishing business as everything moves online and display ad prices race for rock bottom, just as publishers need the funds to build a better website. Then mobile comes along and apps become a necessity, followed by a responsive website—or, better yet, a mobile-first website. Now, publishers have to keep up with the demands of consumers who are not only increasingly mobile, but who are going screenless—using their voices to search for content with Siri, Alexa, and Google Assistant. And, of course, the best monetization methods for voice content aren’t clear yet. Meanwhile, no matter which channel your audience prefers, the content needs to be personalized.

And while all the research indicates that consumers prefer personalized content and are even willing to pay for it, about 30% of online publishers still aren’t implementing it. According to a survey of 200 publishers by Digiday in February, seven out of 10 online publishers personalize the content they deliver to visitors. According to an article by Monojoy Bhattacharjee about the Digiday survey, of the publishers that haven’t yet implemented personalization, Digiday found most “have not done so primarily because of technical (68%) and monetary (58%) constraints.”

And the personalization puzzle just keeps getting more complicated. According to Bhattacharjee, “The New York Times, one of the early adopters, has implemented various personalization strategies, e.g., based on factors such as readers’ geographic location and what content they read. The Boston Globe has used personalization technology to tailor both editorial and marketing content. Hearst Newspapers has used Google’s natural language processing to personalize stories for readers.”

It’s no longer enough to provide one-to-many personalization. The power of AI now makes it possible to offer one-to-one personalization, and as consumers become more used to the level of personalization from Amazon and Facebook, they come to expect it from all the digital brands they interact with—including publishers. In other words, just because you’re providing personalization doesn’t mean you’re meeting your customers’ expectations.

Not coincidentally, AI can help deliver real-time, one-to-one personalization for every reader—but it will cost you. You can see how small publishers would struggle to keep up—and fall further and further behind with each advancement in technology and expectations. But if there was ever a time to get creative and marshal your resources, it’s now. AI isn’t going anywhere, and it has the potential to help save/make you money. You can’t afford to wait.