Skip to main content

How ezbot Works

ezbot is an artificial intelligence system that watches over your website. When a user uses your website or app, ezbot watches them and learns from their behavior. The next time a similar user visits, ezbot uses this knowledge to deliver an experience more likely to lead to a successful outcome. Over time, ezbot optimizes the experience for different types of users.

Example: Users Have Different Preferences Based On Weather
Imagine you have a small souvenir shop with two items: baseball caps and beanies. Beanies are warmer than baseball caps. Both are very popular and people all over the US buy them. After a while, ezbot would notice that people with IP addresses corresponding to locations with warm weather buy more baseball caps, and people in locations with cold weather buy more beanies. Over time, ezbot will suggest beanies to people in colder weather and baseball caps to people in warmer weather

Think of it as a nudge. Some people will still buy beanies when it's hot out or baseball caps when it's cold. The important change is that before ezbot, there were some people in warm climates who saw beanies first and some people in cold climates who saw baseball caps first and left the page. ezbot helps you show the right thing to the right person at the right time.

Serving What Works

Once you've created a project and variations to test, ezbot begins its work. It starts by serving combinations of variations to users who are not in the configurable holdback group, and it monitors their interactions.

As more people use your website or app, ezbot detects patterns that are invisible to humans and learns which combinations work best for users with specific attributes. When a new user lands on your website, ezbot serves them the combination of variations that will maximize your reward strategy.

how it works


ezbot helps your business keep up with changing user preferences.

Your website will not only serve different variations to users with different attributes, but will also adapt to serve different variations at different times. User preferences may change based on the time of day, season, or other factors.

For instance, if you run an ad campaign that attracts users with different preferences than your usual audience, ezbot will automatically notice these changes, adjust its model, and serve the optimal combinations accordingly.

Continuous Learning

ezbot employs a proprietary exploration policy to continuously test all of your variations, even if some are under-performing. Although under-performing combinations will be shown less frequently, this approach allows ezbot to identify variations that may perform well under certain conditions or at a later time.

By continuously learning and adapting, ezbot ensures that your website or app always delivers the best possible experience for your users.