AI-optimized optimizations overview

Updated

Learn how Optimize uses AI to automatically show the best-performing variations to each visitor.

AI-optimized optimizations use machine learning to adaptively deliver the most effective variation to each visitor. Instead of declaring a single winner, Optimize continuously learns from user behavior and promotes the variation that’s most likely to drive engagement — personalized in real-time, at scale.

This type of optimization is ideal when you want to continuously promote top performers without picking a winner, adapt messaging based on live performance, or run an optimization indefinitely to maximize conversions over time.

How to get started

Any optimization that is AI-optimized combines the best of testing and personalization — helping you discover what works best while automatically tailoring the experience for every visitor. Once launched, Optimize uses machine learning to promote the variation most likely to drive engagement for each individual.

To create an AI-optimized optimization, do one of the following from the Optimizations menu:

  • New test > AI-optimized — try out messaging or design ideas
  • New personalization > AI-optimized — adapt content based on audience segments
  • AI-powered suggestions — brainstorm context-aware ideas with help from AI, based on the current page’s content

Regardless of where you start, your optimization continuously learns from engagement data and adjusts traffic to serve the best-performing variation to each visitor. For more control, you can apply an audience to hone your personalized ideas.

How it works

AI-optimized optimizations don’t split traffic evenly or declare a single winner. Instead, they use engagement data to learn which variation performs best for each type of visitor — based on known signals and real-time behavior. As more data comes in, Optimize dynamically shifts traffic to drive the most conversions.

In one AI-optimized optimization, there may be multiple “winning” variations — each shown to the subset of visitors most likely to convert. The AI continuously adapts to match the right variation to the right person.

You can add new variations at any time. As visitor behavior shifts — whether due to new ideas or changing patterns — the AI reevaluates performance and automatically adjusts traffic to prioritize the top-performing variations.

Key benefits

  • Dynamic traffic split — Optimize adjusts traffic in real time based on variation performance
  • No fixed “winner” — the AI promotes whichever variation drives the most engagement for each visitor
  • Flexible management — add or pause variations at any time without restarting the optimization
  • Automatic learning — Optimize continuously adapts as visitor behavior changes

Compare optimization types

AI-optimized test vs. traditional test

Traditional test AI-optimized test
Fixed traffic split Dynamic traffic split based on engagement
Need to stop the optimization to add new variations Can add new variations at any time
One “winner” is declared AI promotes best-performing variation per visitor
Takes longer to reach definitive results Learns and adapts quickly
Ends when a winner is found Runs indefinitely and optimizes continuously

AI-optimized personalization vs. manual personalization

Manual personalization AI-optimized personalization
Manual audience targeting AI targeting based on behavioral signals
Persistent experience per audience Variation shown may differ per visit
Performance reports by segment Performance personalized per visitor
You choose which variation to show AI determines which variation to show
Static experience unless updated Continuously learns and adapts