What Goes Into the AI Report

An AI Report presents the findings from your study as a set of slides you can share with your team. This article explains what a report contains, how it differs by study type, and what determines its content.

HOW A REPORT IS STRUCTURED

A report follows the same arc regardless of study type.

It opens by establishing what was studied: the report title, a short description of the analysis, and the headline facts such as sample size, what was tested and the areas the report covers.

It then presents the findings, working through the data your study collected. Each finding is shown with the figures behind it, usually as a chart with the key numbers called out alongside.

It closes with a strategic summary: an overall verdict on what the study found, and a set of prioritised recommendations.

WHAT THE REPORT DRAWS ON

The report is built entirely from data already in your study. Nothing is introduced that was not collected.

Participant profile. Who took part, broken down by the demographics captured in your study.

Behavioural data. What participants did during the study.

Question responses. How participants answered, with the distribution shown as a chart and the key figures summarised beside it.

Qualitative feedback. What participants said in their own words, organised into what worked and what needs improvement.

Recommendations. An overall verdict and prioritised actions, drawn from the patterns across everything above.

WHAT DIFFERS BY STUDY TYPE

The findings section reflects what the method actually produced.

Prototype and task-based studies. Task completion, drop-offs, average time and journey success rates, screen-level usability scores, and the responses given alongside the task.

Qualitative studies and AI interviews. Themes identified across sessions, each carrying the number of sessions it appeared in so a pattern across participants is distinguishable from a single reaction. Verbatim quotes evidence the findings. Emotion and behavioural metrics from the recordings sit alongside them.

AI Creative Insights. The Decode Score and the individual metrics, each read against the category benchmark. Time to Discover and Attention per area of interest where areas were marked. Where a synthetic audience was applied, how the creative performed for each persona.

Reports present data and findings rather than visuals, so attention maps stay in the results view. Where a finding rests on where attention landed, open the results alongside the report.

WHAT DETERMINES THE CONTENT

Three things you set at generation shape what you get back.

The objective. This is the main determinant. A report generated with an objective about attention and clarity gives weight to those measures. The same study with an objective about emotional response and improvement recommendations draws on different parts of the data and reaches different conclusions. Write the objective as the question you need the report to answer.

The Objective Template. This sets the lens. Generic generates from your objective as written. A specific template such as Creative Performance or Packaging Test and Shelf Performance gives more weight to the metrics relevant to that area.

The format. An Outline Report gives the key findings in summary form. A Detailed Report covers the data in depth with the underlying figures. An outline suits a stakeholder update. A detailed report suits a working session where the numbers will be questioned.

HOW FINDINGS ARE QUALIFIED

Where a finding rests on a small number of responses, the report states this on the slide alongside the number it is based on. The closing section carries the same qualification for the study as a whole.

Read these before presenting. They tell you which findings are firm enough to act on and which are directional.

CHANGING WHAT A REPORT CONTAINS

Reports cannot be edited after generation. To change one, generate a new report with a revised objective.

Each report appears as its own card on the Report tab, so earlier versions remain available alongside the new one.