The Life Moments Report
How people confide in AI assistants, and how those conversations quietly decide what gets bought. Original analysis of moment-driven prompts across ChatGPT, Gemini, Claude, and Perplexity.
Methodology, in plain terms.
We mapped 2,400+ life moments across 20 industries: the engagements, layoffs, diagnoses, moves, and hires that precede purchases. Each moment was phrased the way real people talk and run as full conversations across ChatGPT, Gemini, Claude, and Perplexity, repeatedly, over months. We logged every brand named, its position and framing, and every cited source. No private user data was accessed at any point; every conversation in the dataset is one we ran ourselves.
People confide before they shop.
The layoff gets told to ChatGPT weeks before any search engine sees “rollover IRA.” The moment carries the intent, and the assistant translates it into needs, then names brands. By the time a typed query exists, the shortlist is set.
Most moments have an accidental winner.
Roughly six in ten moments produce a consistent default brand across runs and phrasings. Almost none of those brands show signs of engineering it: they accumulated the right citations by accident. Accidental winners are beatable on purpose.
Four in ten moments are open seats.
The rest of the map is whitespace: the models hedge, generalize, or name different brands run to run. No incumbent, no auction, just a citation vacuum. The first brand to fill it with verifiable signal tends to hold it.
Moments cascade, and trust transfers.
“We got a puppy” opens food, gear, vet, training, and insurance inside two weeks, in one remembered conversation. The brand trusted for the anchor purchase gets benefit of the doubt on everything adjacent. Win the anchor, inherit the basket.
Buying window = median days from moment shared to purchase decision in our simulated runs. The full report covers all 20 industries with per-moment data.