Marketers are used to the concept of split-testing possible content and communication changes to determine a winning combination. The problem with a standard A/B split test is that it can still leave many of your customers with a sub-optimal experience because you’re assuming that all of your customers fall into the same demographic.

And in today’s marketing, generalization is the enemy.

Consumers now want relevant communications and messages that are tailored to them. In fact, 77.5 percent of “digital natives” want you to give them a truly personalized experience, both on your website and within messages, and that percentage remains high for other age groups, too.

To try to solve this problem of serving content changes to everyone indiscriminately, FollowAnalytics has announced today its new Mobile Optimization Suite, which includes Optimize AI, a new split-testing technology powered by machine learning.

Mobile Optimization Suite includes various new tools to help marketers get the most from iterating, improving, and measuring the success of changes to websites, communications, in-app messages, and push notifications.

It automatically determines the sample size for your test to ensure statistically significant results, while making sure the experiment finishes as early as possible. But most importantly, the suite includes segment-based optimization that allows marketers to run different tests and serve relevant content to audiences based on their personal needs.

“The segments used by Optimize AI are of two types,” Samir Addamine, founder and chairman at FollowAnalytics, told me. “Segments can be defined by the customer, based on mobile data and any external data coming from their CRM or other system of record. And segments can be based on activity in an app. These segments allow you to get results out of the box, even before you have defined any custom segment.”

Split-testing different content based on segments means you can optimize your app to provide an ever-improving personalized experience for the consumer. So how does machine learning power this solution?

“Machine learning allows us to improve results over time,” Addamine said. “The intelligence layer allows us to find the right message for each user targeted by a campaign. Learning allows us to leverage all past campaigns to better assign variants to targeted users. By categorizing your variants along taxonomies, you help the learning algorithm find relationships between distinct campaigns.”

This means you can write content for each type of consumer using your app, and the system will learn what is best for each variant, optimizing per segment rather than expecting everyone to react positively to a wholesale change. In a world where personalization is expected, technologies like this can help drive increased relevance.

Along with these new split-testing capabilities, FollowAnalytics has enhanced its message-building tools to provide greater flexibility. These tools allow you to create push notifications and in-app messages easily, and the suite offers a template library to help you build pop-ups, banners, and native alerts. You can also create custom HTML5 messages.

Addamine thinks AI and machine learning will be an increasingly important part of the marketer’s armory for 2017.

“AI and machine learning will help reach the holy grail of any marketer: the self-driving campaign, my vision since I started FollowAnalytics in 2013,” Addamine said. “Thanks to AI/ML technology, we’ll start seeing very smart and accurate recommendations, creating campaigns that will lead to more retention and engagement. Building more bridges between the CRMs, ERPs, and marketing tools will also catalyze the quality of the predictions.”

FollowAnalytics Mobile Optimization Suite, which includes Optimize AI, is available from today.