What is attention prediction in advertising?

Attention prediction in advertising uses AI models trained on human visual behaviour to forecast where people will look before they consciously process an ad. It enables brands to optimise creative for maximum impact without relying on traditional eye tracking studies.

Attention prediction in advertising refers to the use of artificial intelligence (AI) to simulate and predict human visual attention. By analysing elements such as contrast, colour, composition and object placement, these models estimate which areas of an image or video are most likely to attract attention within the first few seconds of viewing.

The technology is rooted in decades of neuroscience and vision science research, particularly studies into how the human brain processes visual stimuli before conscious awareness.

This includes saliency modelling, which maps how certain features naturally stand out to the human eye. Modern AI systems build on this foundation using deep learning trained on large datasets of human gaze behaviour.

Unlike traditional eye tracking, which requires physical participants and specialist hardware, attention prediction can be run instantly and at scale. This makes it far more efficient for marketers testing multiple creative variations across channels such as social media, display advertising and ecommerce.

Attention prediction matters because attention is a limited resource in today’s digital environment. Ads that fail to capture attention within the first second are often ignored. By predicting attention before launch, brands can design more effective creative that improves engagement, recall and overall performance.

How Dragonfly AI uses attention prediction

Dragonfly AI applies attention prediction to help suppliers and brands optimise creative assets before they go live. Using AI models developed in collaboration with researchers from Queen Mary University of London, the platform simulates human attention to identify which elements of an ad will be seen first.

Users receive visual heatmaps and Attention scores that highlight high and low performing areas of their creative. This allows marketers to adjust layouts, reposition branding or refine messaging to improve visibility.

For example, a brand running social media campaigns can test multiple versions of an ad to ensure key elements such as logos, product imagery or calls to action are seen immediately. This reduces wasted media spend and increases campaign effectiveness.

Attention prediction in the UK market

The UK has been at the forefront of attention research, with strong academic and industry collaboration driving innovation in this area. Dragonfly AI’s technology is rooted in research conducted at Queen Mary University of London, a leading institution in vision science and AI.

Industry bodies such as the UK Attention Council have highlighted the growing importance of attention metrics in evaluating advertising effectiveness. As the attention economy becomes more central to marketing strategy, UK brands are increasingly adopting tools that go beyond impressions and clicks.

In a highly competitive media landscape, UK advertisers are under pressure to demonstrate measurable effectiveness. Attention prediction provides a scalable and evidence-based approach that aligns with the UK’s strong culture of effectiveness, reflected in initiatives such as the IPA Effectiveness Awards.