How it works

In short

Foresight checks and prepares your creative, runs an attention model and an emotion model in parallel, turns their predictions into 0–100 scores against reference creatives, and then writes AI insights. Images are analysed as a single view, videos moment by moment.

Every Foresight analysis follows the same path: your creative is checked and prepared, two prediction models read it, their predictions become scores, and AI insights explain what the scores mean. This section explains each part in plain language.

From upload to insights: both predictions run in parallel, then become scores positioned against real creatives.

• The analysis flow: the journey from upload to results, step by step, and which of your choices affects which step.
• The emotion model: how Foresight predicts which emotions your creative is likely to evoke, and what it learned from.
• The attention model: how Foresight predicts where people are likely to look, and how visual complexity is estimated.
• How images are analysed: a single moment of viewing, with gaze plot, attention build-up and areas of interest.
• How videos are analysed: moment-by-moment predictions, timelines and scenes, and why video scores summarise the whole video.

Two models, one result

The attention model sees only pixels; the emotion model sees your original creative and also uses a short description of the creative's content as context. Both learned from real human data, and both describe a typical viewer. See Predicted vs measured.

Images and videos go through the same steps and get the same KPIs on the same 0–100 scale, but they are compared with different benchmarks, and every score is positioned against creatives of the same media type (see how scores are positioned). That is why Foresight keeps them apart: compare images with images and videos with videos.