The challenge
- User signal extraction existed only in Proctor — used for monitoring, not for learning personalization or Construct enrichment
- Available devices (wearables) not integrated into Construct alongside existing product data;
- Challenges with deployment across jurisdictions, data availability and privacy constraints.
Goal:
- Use available devices to enrich Construct for two purposes:
- (1) scientific recommendations to improve software quality;
- (2) increase user attachment and retention.
The Solution
- Enrichment of product capabilities (Learn, Assess, Schedule, Groups) by integrating wearable data into the construct;
- Personalisation: selection of highest-value signals for user context modelling, adapted to the individual and to their available devices;
- Verified fusion and policy algorithms for context-aware event generation and additional feedback opportunities;
- Support for explainability, privacy, autonomy, robustness, and edge–cloud optimization.
The project includes 2 research streams:
- Stream 1: Personalization from existing C>T product data fusion.
- Stream 2: Personalization from wearables + product data fusion (focus/energy, circadian, stress, readiness).
Impact
- Context-aware personalization;
- Bi-directional feedback loop: sensing user state and sending signals back to via their devices.
- Validated multi-source data fusion under privacy and robustness constraints;
- Scalability up to 300m users
Industrial Directions
- Digital collaboration platforms.
- Learning, assessment, scheduling, and group coordination systems.
- Productivity environments leveraging wearable-based personalization;
- Research platform
- Digital Platforms & Transformation, Education, Enterprise Software, Human Performance & Workplace, Human–Computer Interaction, Knowledge & Intelligence Systems, Learning Systems, Wearable Technology
