Discover is an AI-powered way to explore qualitative research inside a Lookback project.
It is designed to help researchers and teams navigate, question, and revisit their data - not to replace human judgment or interpretation.
Discover works as an agent inside the project, grounded in recordings, transcripts, notes, and findings.
WHEN TO READ THIS
Read this article if you are:
• working with many sessions or long recordings
• trying to understand patterns across a project
• involving stakeholders in exploration
• curious about how AI fits into qualitative sense-making
WHAT DISCOVER IS FOR
Discover helps answer questions such as:
• What are people confused about across sessions?
• Where do we see similar reactions or behaviors?
• What evidence relates to a specific topic or theme?
• What did we learn about a concept we did not initially focus on?
Discover is especially useful when insight emerges over time, rather than in a single moment.
DISCOVER AND HUMAN JUDGMENT
Discover does not make decisions or conclusions.
Instead, it:
• summarizes
• points to relevant evidence
• highlights patterns
• surfaces quotes and moments
Researchers remain responsible for:
• interpretation
• prioritization
• deciding what matters
This is a deliberate human-in-the-loop design choice.
THE THREE THINGS RESEARCHERS NEED TO FIND
In qualitative research, teams typically need to find three kinds of things:
Things they knew they were looking for before research began
Things they recognized as important when they encountered them
Things they did not recognize as important until later
Discover is especially valuable for the third case.
It helps teams realize:
• that something has come up repeatedly
• that a pattern exists across time or segments
• that earlier observations take on new meaning
DISCOVER AND EVIDENCE
Discover always points back to evidence.
Its responses include:
• references to sessions
• quotes from transcripts
• links to specific video moments
This keeps exploration grounded and verifiable.
Discover reduces the need to rewatch entire sessions, while still preserving the principle that seeing is believing.
WORKING WITH STAKEHOLDERS
Discover can be useful for collaborative exploration.
Stakeholders can:
• ask questions in natural language
• explore areas relevant to their work
• connect their perspective to real user evidence
This often surfaces insights that researchers alone might not prioritize.
WHAT DISCOVER IS NOT
Discover is flexible and capable, but it is important to understand its role.
It is not:
• a replacement for watching important moments
• an automatic insight generator
• a reporting or presentation tool
Think of Discover as an exploratory partner that helps you ask better questions of your data.
WHAT TO READ NEXT
• Roles and Access in Lookback – to understand who can use Discover
• Working With Recordings, Findings, and AI – for the broader analysis workflow
• Highlight Reels and Sharing – for communicating insights