Press

For journalists
and researchers.

Key facts, verifiable claims, researcher background, and contact for media enquiries about the ABSD framework research.

Key claims

What this research asserts - and what supports each claim.

→
73% of companies cannot use behavioural data beyond basic segmentation. This gap has a structural cause: the absence of a framework connecting observable actions to psychological motivation.
Forrester Research, 2022 · The state of customer analytics
→
Netflix processes 500 billion events per day but cannot determine whether a user's pause indicates anticipation, confusion, or discomfort. The data exists. The interpretation framework does not.
Netflix Engineering, 2015 · Covington, Adams & Sargin, RecSys 2016
→
State detection from interaction data achieves 70–79% accuracy in controlled settings using affective computing methods. Most deployed AI systems do not attempt it.
Calvo & D'Mello, IEEE Transactions on Affective Computing, 2010
→
Four established research traditions address this problem - Self-Determination Theory, affective computing, JITAI research, and learning analytics - but the literature synthesis found minimal cross-citation between them.
Zawacki-Richter et al., 2019 · Literature synthesis conducted for this study
→
78–84% accuracy predicting student dropout from behavioural pattern data alone - without any state inference. The signals that would enable deeper understanding are already present in standard system logs.
Li, Baker & Warschauer, Internet and Higher Education, 2020 · Kuzilek et al., 2017
Research facts

At a glance.

Title
Predicting the Why: The ABSD Framework for Inferring Psychological States from Behavioural Signals in AI Systems
Type
Applied Research Project · SP Jain School of Global Management · Executive MBA
Researcher
Arunamirtharaj Chandrasekar · Enterprise technology professional · 20+ years APAC cloud consulting
Mentors
Himmat Singh (academic) · Akshit Sharma, Brained.app (industry)
Methodology
Qualitative · Expert interviews (semi-structured, n=5) · Practitioner survey · Literature synthesis (53 sources)
References
53 sources across 4 traditions: SDT, affective computing, JITAI, learning analytics, plus supporting CX, AI, and methods references
Access
Open access · Full paper readable at predictwhy.com/paper · No embargo · No paywall
Important caveats

What this research is not.

This is not a technology product. ABSD is a conceptual framework - a map of the problem and a proposed architecture. It does not include software, an algorithm, or a deployed system.

This is not a surveillance proposal. The framework explicitly addresses the ethical requirements of transparency and consent for any system applying its principles.

This is not a claim that AI can read minds. The confidence gradient is explicit in the framework: state inference is probabilistic, not deterministic. Drive inference is longitudinal, not instantaneous.

The practitioner validation is qualitative. The research demonstrates that practitioners recognise the problem ABSD describes and find the framework useful. It does not measure implementation outcomes.

Media contact

Get in touch.

For interview requests, fact-checking, or questions about the research, contact the researcher directly. Response time is typically 24 to 48 hours.

Researcher
Arunamirtharaj Chandrasekar
SP Jain EMBA · Singapore · predictwhy.com
arunchandru@outlook.com
Send a message →