The AI narrative never gets to replace the real numbers.
A lot of AI reporting tools show you a paragraph and hide the data behind it. Brand Perception does the opposite: the actual counts and percentages are always visible, and the AI summary sits on top of them, not instead of them.
Why this matters more than it sounds
When you present results to a client, "the AI thinks respondents preferred option B" is not a defensible statement on its own. "68% of 140 respondents preferred option B, and they were 40% faster to decide than on option A" is. The difference is whether the actual data is right there for anyone to check, or whether you're asking people to trust a black box.
What stays visible, always
- Exact response counts and percentages, per question, per campaign
- A full CSV export of raw responses, available on every plan
- Response timing and switch-count data, not just the final answer
What the AI narrative actually does
It reads the same numbers you can see, works out what kind of decision the campaign is testing, and writes the verdict that decision needs. It's a presentation layer on top of real data, generated once when a campaign closes so the numbers behind it always match exactly what was collected. If data used in a report is later erased, the report is marked stale rather than quietly rewritten, and regenerating it is always a deliberate choice you make.
About the reporting
Can I export the raw numbers?
Yes, every campaign has a CSV export of raw responses alongside the results view and the AI report.
What happens if the AI gets a summary wrong?
You can check it against the real counts sitting right underneath, since the narrative never replaces them. If the underlying data changes (through erasure, for example) the report is marked stale rather than silently rewritten.
Is this just a chart, or the actual data?
Both. The results view shows exact response counts and percentages per question, not rounded or smoothed, and that same data is what the AI narrative is built from.