In Lookback, understanding does not come from summaries or dashboards.
It comes from evidence.
That evidence begins with sessions and becomes meaningful through findings — short, timestamped moments that capture something worth paying attention to.
This article explains how that transformation works, and why Lookback is designed this way.
Sessions are raw material
Sessions capture what actually happened:
what participants did
what they said
where they hesitated, struggled, or changed course
the context in which behavior occurred
Sessions are necessary, but they are not yet usable insight.
Qualitative research requires selection, interpretation, and comparison - work that starts live, happens during, and continues after sessions.
Findings are the primary unit of evidence
A Finding is a short, timestamped video clip taken from a session.
Findings represent moments that matter:
a point of confusion
a workaround
a statement that reveals intent
a contradiction between what someone says and does
Findings are not conclusions.
They are evidence.
Each finding stays directly connected to the original session and moment in time.
Why findings are video-based
Lookback treats video as central because qualitative insight often depends on:
tone
timing
hesitation
body language
context
Video preserves ambiguity and nuance in a way summaries cannot.
This allows stakeholders and researchers to:
see the same evidence
interpret it together
challenge assumptions without relying on second-hand descriptions
Findings emerge in three ways
Findings can be created through different forms of attention:
Things researchers knew they were looking for
Researchers define goals in advance, often informed by stakeholder input. These goals help focus attention during and after sessions, and provide context for Eureka.
Things researchers notice as they happen
During live or replayed sessions, researchers may recognize important moments and capture them immediately. Lookback makes it easy to capture those moments while under cognitive load.
Things only recognized later
Some patterns only emerge when sessions are compared, revisited, or viewed by others. Lookback supports returning to evidence without losing context.
AI assists in all three cases - but does not replace judgment.
Themes group related findings
A Theme groups related findings across sessions.
Themes help researchers:
see patterns over time
compare evidence across participants
understand variation and consistency
Themes do not flatten evidence. Each finding remains visible and traceable.
Themes support synthesis - not abstraction away from evidence.
Reels are for storytelling, not stronger evidence
Reels are collections of findings assembled to:
illustrate a pattern
tell a story
communicate insight efficiently
Reels do not create new evidence.
A single finding can be sufficient on its own. Reels simply provide a way to present multiple findings together when that is helpful.
The role of AI in analysis
Lookback’s AI is designed to reduce cognitive load, not to automate conclusions.
AI can help:
surface moments aligned with defined goals
highlight patterns across sessions
reduce the amount of video that must be manually reviewed
AI does not decide what is important.
It assists researchers in staying close to evidence while working more efficiently.
Human judgment remains central.
Why this evidence model matters
By keeping findings:
timestamped
video-based
tied to sessions
visible to stakeholders
Lookback supports:
transparency
shared understanding
evidence-based decision-making
This makes it harder to over-summarize, over-generalize, or lose nuance - and easier to reason together about what is actually happening. This sets the tone for great collaboration.
How this fits the bigger picture
Sessions capture what happened
Findings mark what matters
Themes reveal patterns
Reels help communicate insight
All of it remains grounded in observable evidence.
What to explore next
To go deeper:
Learn how research roles interact with evidence differently
Explore how stakeholders collaborate around findings
See how AI supports sense-making without bypassing judgment