Leverage the best science to optimize ad creative

Our models are inspired from state-of-the-art research coming from the best labs at Harvard and MIT. They take millions of human reactions to content and learn what makes an impact on your audience. They then become your co-pilot to optimize every creative decision.

How AI changes the game

High accuracy

Trained with massive data

De-risk from the start

Don’t wait until the end to test and correct

Results take seconds

Never delay a launch

Fully scalable

Test all ads and every creative decision

Why predict cognitive impact

Cognitive factors like memorability are not intuitive to the human eye. When given two images or videos, a person’s prediction of which one will be the most impactful is around 52%, the same as flipping a coin.

Every misguided decision based on guesswork costs you money. You need a great deal of impressions to build a brand or to generate clicks.

Achieve more from every impression

Saliency

Build ads that attract the attention of audiences in each context. Redirect their attention towards your key elements.

Ad recall

Stay in people’s minds and build trust, perceptions of quality and brand loyalty.

Conversion

Convert your audience with fewer impressions. Leverage all the data of what worked and what didn’t work for you in the past.

How it works

13.1 Million
reactions to image and video ads
Step 1

We collect millions of human reactions to content

Scenes
People ethnicities, ages
Text in ad and description
Saliency of each element
Call-To-Action
Evoked emotions
Advantages & discounts
Object motion
Intra-Image Context & Surprise
Colors & brightness
Product & logo presence and saliency
Video length
Difficulty to understand
Distinctiveness in category
Scene complexity
Camera motion
Ad saliency in context
Actions
Scenes
People ethnicities, ages
Text in ad and description
Saliency of each element
Call-To-Action
Evoked emotions
Advantages & discounts
Object motion
Intra-Image Context & Surprise
Colors & brightness
Product & logo presence and saliency
Video length
Difficulty to understand
Distinctiveness in category
Scene complexity
Camera motion
Ad saliency in context
Actions
Step 2

We train deep-learning models to predict impact

97%
Accuracy predicting saliency
0,03
Average absolute error predicting memorability
8.6/10
Times our AI selects the best asset among alternatives
Step 3

We test these models against human behavior to evaluate their accuracy

Increase your
ROAS now