90-second read
Everything that matters.
Nothing that doesn't.
The complete ABSD research condensed to one page. Read this first. Follow the links when you want more depth.
Estimated reading time: 90 seconds
The problem
AI systems track what users do. They miss why.
73% of companies cannot use behavioural data for anything beyond basic segmentation (Forrester, 2022). Companies that personalise well generate 40% more revenue than average (McKinsey, 2021). The gap between those two numbers has never been formally named. This research names it.
500B
Events per day · Netflix
73%
Stuck at basic segmentation
40%
Revenue uplift from personalisation
The ABSD framework
Four layers. Each one sits beneath the last.
Drive generates State. State produces Behaviour. Behaviour manifests as Action. AI systems observe Action and must infer upward. Most never try.
L4
Drive
Why. Competence, autonomy, relatedness.
L3
State
What users feel. Confusion, flow, frustration.
L2
Behaviour
Patterns over time. Hesitation, regression.
L1
Action
Observable. What most AI captures today.
Confidence degrades as you go deeper. Action is observable at 96% fidelity. State inference achieves 70–79% accuracy in controlled settings. Drive inference is longitudinal. The framework is honest about where inference stops being reliable.
Action
96%
Behaviour
82%
State
70–79%
Drive
Longitudinal
Four traditions
Each one solves part of the problem. None of them alone.
ABSD synthesises four established research streams that had never been connected. Each tradition provides a specific methodology for a specific layer transition.
Self-Determination Theory
Grounds the Drive layer. Competence, autonomy, relatedness.
Affective Computing
Grounds the State layer. 70–79% accuracy from interaction data.
JITAI Research
Grounds the Behaviour layer. Vulnerability and receptivity windows.
Learning Analytics
Grounds the Action layer. 78–84% dropout prediction from patterns.
Seven signals
The bridge between layers. All already in your data.
Hesitation
Pause before action. Indicates uncertainty or low confidence.
Thrashing
Rapid switching without completing. Overwhelm or decision paralysis.
Regression
Returning to earlier content. Confusion or need for consolidation.
Acceleration
Speeding through content. Boredom, overconfidence, or disengagement.
Persistence
Continuing after failure. Strong competence drive. Do not interrupt.
Avoidance
Skipping, early exit. Often anxiety, not disinterest. Clearest churn predictor.
Flow
Steady progress, low errors. The optimal state. Correct response: leave it alone.
The research
Open. Live. Practitioner-led.
53 references across 6 traditions. Live survey and expert interviews. Two case studies - Netflix and Khan Academy - analysed through the ABSD lens. The research question: do practitioners recognise the signal gap in their own systems, and does the framework give them vocabulary to act on it?
14
Survey responses
3
Interviews done
50
References mapped
Try it interactively
Pick a domain. Watch a signal travel through all four layers.
The ABSD Signal Tracer shows exactly where current AI systems stop inferring — and what ABSD-aware inference looks like in EdTech, CX, Healthcare, and Enterprise SaaS.
Trace a signal →