How the Product Evolved
The product evolves rapidly because the team watches every session and changes things when something causes confusion. Flows get replaced, buttons move, entire features disappear: all responses to what is observed in real user sessions.
From answer engine to AI Twins
Section titled “From answer engine to AI Twins”When Consumr.AI started, the team built an answer/conversation engine. The idea was simple: give users access to a report, let them ask questions, get answers. Conversation engines were just becoming viable, and it seemed like the right approach.
It failed in practice. Experienced marketing veterans sat in front of the product and had no idea what to ask. They typed random questions. The problem was not the answers; it was that users did not know what questions to ask to get value.
The team replaced the answer engine with AI Twins. Instead of asking a report, you ask a twin, a persona that impersonates a consumer cohort. This gave users something to talk to, which made the right questions much more natural. You don’t have to know what to ask a dataset; you know how to interview a person.
From “attach 3 reports yourself” to a guided flow
Section titled “From “attach 3 reports yourself” to a guided flow”When the twin build first shipped, the process was simple: attach three reports to a twin, you’re done. Users consistently did not know which reports to attach or why. The product team watched session recordings and saw the confusion play out over and over again.
The team added a guided flow that walks users through segment definition, brief generation, input selection, and report creation in sequence. The UI keeps changing because this process is still being refined based on what users find natural or confusing.
From “stages of conversion” to the mindset metric
Section titled “From “stages of conversion” to the mindset metric”Early versions of the respondent distribution screen showed a stages of conversion metric, a way to classify where in the funnel a respondent sits. This metric was too rigid to express in a fluid distribution interface, so the team began moving to the mindset metric, which is better suited to representing the range of attitudes within a cohort. That migration is still in progress; the older stages-of-conversion view may still appear in places until it is finished.
Why it keeps changing, and what that means for you
Section titled “Why it keeps changing, and what that means for you”The team watches every session that goes on in the product and knows exactly what users find comfortable and what they do not. When something causes confusion, the team changes it. Fast. The ops team and business team sometimes feel like changes come too frequently; the frequency is evidence that the product is being actively improved based on real observed behavior, not intuition.
As someone working inside the product, expect flows to evolve. When something looks different from what you were shown, check the release notes at resource.consumr.ai or ask in the team channel. Do not assume the old way still works.