predictwhy.com ×
Tap to begin · 13 screens
Predicting
the Why.
A guided walkthrough of the ABSD framework research. Tap the right side to advance. Tap the left edge to go back. Hold to pause.
The problem
73%
of companies stuck at basic segmentation
40%
more revenue from good personalisation
500B
events/day · Netflix · still Layer 1 only
The data exists.
The interpretation doesn't.
Core claim
AI systems predict what users will do.
Not why they're doing it.
That is the gap. It has a structure. It has a name. ABSD. Four layers from the observable surface to the motivational core.
The ABSD framework · four layers
Drive
Why · L4 · longitudinal
State
What users feel · L3 · 70–79%
Behaviour
Patterns over time · L2 · 82%
Action
Observable · where AI sits today · L1 · 96%
Drive generates State → State produces Behaviour → Behaviour manifests as Action. AI must infer upward.
Confidence gradient
Deeper inference.
Lower certainty.
Action96%
Behaviour82%
State70–79%
DriveLongitudinal
Four traditions · never previously connected
Self-Determination Theory
Drive layer
Affective Computing
State layer · 70–79%
JITAI Research
Behaviour layer
Learning Analytics
Action layer · 78–84%
Each tradition solves one layer. ABSD is the first framework to connect all four.
Seven signals · already in your data
Hesitation
Pause before action · uncertainty
Thrashing
Rapid switching · overwhelm
Regression
Returns to earlier content · consolidation
Acceleration
Speeding through · boredom
Persistence
Continues after failure · high drive
Avoidance
Early exit · anxiety
Flow
Steady progress · optimal state · don't interrupt
Critical qualification
Same signal.
Different context.
Different meaning.
A 30-second pause before a complex financial trade is prudence. Before a simple form field it is confusion. Three modifiers determine which: domain, user history, and temporal context. Apply all three before acting on any signal.
Case study · CX · Netflix
500 billion events
per day. Layer 1 only.
When you pause a film, Netflix logs the event. It cannot determine whether that pause was anticipation, confusion, or discomfort. Three possible states. One data point. The signals for Layer 3 inference are in the data. The framework to read them is not.
Case study · EdTech · Khan Academy
Mastery tracked.
Not understood.
A student who answers correctly after 45 seconds of hesitation receives the same credit as one who answers in 4 seconds. VanLehn (2011): AI tutors underperform human tutors precisely because they cannot read the psychological state behind the answer. That gap remains accurate today.
The meta-insight
Right now, you are being observed through the framework you are learning about.
The survey captures how you answer - hesitation, pace, answer changes - not just what you say. The closing card reflects your ABSD layer profile back at you. The thesis demonstrated, not just described.
Open research · live now
50 references.
Live survey.
Always open.
50
References · 4 traditions
14+
Survey responses
5
Expert interviews
SP Jain EMBA · Applied Research Project. Full paper always readable - no login, no paywall.
Your next step
Does this resonate
with your experience?
Your answers and your behavioural patterns as you answer are both part of the research.