Collaborated with humans from
← All work
Case studySearchAIUser researchwith IPSOS, for Google
How people really search
Six cities, four generations, about 70 search journeys, as AI arrived.
- For
- Google Search · via IPSOS
- Role
- Primary researcher & moderator
- When
- 2023
- Participants
- 24 · ages 15 to 60
Query logs show what people typed. They never show why someone closed the tab and opened a chatbot instead. So we went and sat with the people.
Step 1 · In-depth interviews across India
Frequent searches, rare ones, and every app in between.
Schematic, not to scale.
Step 2 · Jobs-to-be-done synthesis
Explorers and finishers.
Mindset
{{band.mind}}
A search lasts
{{band.time}}
To finish, they
{{band.finish}}
Phone first. Journeys start on a phone, then move to a laptop when visuals matter.
Social sparks it. A post gives just enough to make people curious.
Google first. For anything unfamiliar, people still start here.
Step 3 · The signal that mattered
When people can't phrase it, they ask an AI.
Some participants tried prompt after prompt to get Google to understand them. Others had already switched to chatbots, because they could pull a clear answer out of a messy question. Study and homework searches showed the most AI already.
that thing where plants make food from sun?? for class
Ten blue links. Try again with better keywords.
You mean photosynthesis, here's how it works, step by step.
Illustrative exchange, written to show the pattern.
Young people weren't leaving for novelty. They wanted a partner for half-formed questions.
What it fed
~70
search journeys mapped
4
AI opportunity areas briefed
4B+
people use the product it informed