How Sentient Ascend Works

In ecommerce, designing user experiences (i.e. web pages and interactions) that convert as many users as possible from casual browsers to paying customers is an important goal. While there are some well-known design principles, including simplicity and consistency, there are also often unexpected interactions between elements of the page that determine how well it converts. The same element may work well in one context but not in others—it is often hard to predict the result, and even harder to decide how to improve a given page.

An entire industry has emerged to tackle these challenges; it’s called conversion rate optimization, or conversion science. The standard method most practitioners use is A/B testing, i.e., designing two different version of the same page, showing them to different users, and collecting statistics on how well they each convert. This process allows incorporating human knowledge about the domain and conversion optimization into the design, and then testing their effect. After observing the results, new designs can be compared and gradually improved. The A/B testing process is difficult and time-consuming: Only a very small fraction of page designs can be tested in this way, and subtle interactions in the design may simply be missed completely.

Sentient Ascend is a disruptive technology that changes all that. Ascend is an AI agent that, based on evolutionary computation technology, automatically generates web-page candidates to be tested from the ideas marketers themselves create. Those variables can be pretty much anything on a site, from small changes like button color and font weight to whole-scale messaging and design ideas. What’s more, they can all be on a single page or even across multiple pages in a funnel–something that’s currently impossible with industry-leading tools. Ascend searches for the most successful ideas in the vast space of possible combinations of the values its users want to try.

And this space can be very large.


Ascend can handle this intelligently because it uses evolutionary computation. Each page is represented as a genome, as shown for two example pages in the figure below (left side). Simulated genetic operators such as crossover (recombination of the elements in the two genomes; middle) and mutation (randomly changing one element in the offspring; right side) are then performed. If the parent genomes are chosen among those that convert well, then some of their offspring genomes are likely to perform well too–even better than their parents. Each page needs to be tested only to the extent that it is possible to decide whether it is promising, i.e. whether it should serve as a parent for the next generation, or should be discarded.


Therefore, thousands of page designs can be tested in a short time, which is impossible through A/B testing. On the other hand, through evolutionary search, it is sufficient to test only thousands (out of millions) to find the best ones. Ascend learns over time which combinations of elements are effective, and gradually focuses the search around the most promising designs—as a matter of fact, at Sentient we have solved problems with spaces as large as 2^2^70 (approximately 10^(10^20)) with similar techniques in the past.


All this means Ascend can discover designs that convert better than those designed by humans, often because it finds unexpected interactions between elements. For instance, Ascend may find that the button needs to be green, but only when it is transparent and the header is in small font and the header text is aligned. Such interactions often do exist, and they can be very difficult to find. Ascend makes this discovery process automatic, based on computing power and AI instead of extensive human effort. With Ascend, it is thus possible to optimize conversions better and at a much larger scale than before–and keep optimizing them as the ecommerce conditions change.

In other words, Sentient Ascend replaces A/B with AI. It’s an automated system for massively multivariate™ conversion optimization, capable of testing vastly more ideas in shortened time frames. It finds the subtle combinations of variables that lead to conversion increases. And because marketers can spend more time trying ideas and less time doing statistics, Ascend gives them the freedom they need to really make a difference.

Learn more at

2 thoughts on “How Sentient Ascend Works

  1. This is pretty Awesome!! Keep up the good work! I will be doing an internship this summer with a NLG company and I would love to keep listening AI news and companies like you! do you have a newsletter?

    1. Thanks, Youssef! We have added you to our newsletter distribution list, and wish you the best with your internship.

Comments are closed.