Innovation & Research
The research that
ProxiShop is built on
A systematic review of 47 peer-reviewed papers across five academic themes, grounded in Design Science Research methodology and the latest human-AI collaboration frameworks.
The Primary Research Gap
What no existing platform has done
Confirmed across 47 papers spanning VUI accessibility, live commerce, conversational AI, and human-AI collaboration.
"No existing system simultaneously combines (1) real time voice command interaction, (2) a live human shopping handler in a physical store, (3) AI product narration, and (4) accessible design principles in a single integrated live commerce platform. No study either has examined the end to end shopping experience of visually impaired users within a live, human and AI mediated commerce session."
ProxiShop Research Gap · Systematic Review 2026
Literature Review
Five interconnected research themes
The academic foundation for every ProxiShop design decision.
Visually Impaired Shopping Experience
McKinsey (2023), Nicoson & Ha-Brookshire (2025), Pal et al. (2024) confirm pervasive digital barriers. Screen readers fail at transactional e-commerce. Human mediation is required.
Voice User Interface Accessibility
Jindam et al. (2023) identify failed STT as the largest VUI barrier, directly informing ProxiShop's MicCoordinator architecture. Deshmukh & Chalmeta (2024) map VUI usability dimensions.
Live Streaming Commerce
Gao et al. (2024) demonstrate human broadcasters generate parasocial trust AI cannot replicate. Chen & Li (2024) show AI virtual streamers fail on professionalism, empathy, and responsiveness.
Conversational AI in Retail
Ferraro et al. (2024) identify six paradoxes of GenAI in customer service. ProxiShop's human handler model structurally resolves the empathy and quality paradoxes while managing residual AI risks.
Human and AI Collaboration & Inclusive Design
Xiao & Holloway (2026) propose the Three Layer Framework for Disability Centred Human Agent Collaboration, adopted as ProxiShop's primary theoretical anchor. Yu et al. (2024) examine the closest analogue in remote sighted assistance.
Theoretical Framework
The three-layer collaboration model
Adapted from Xiao & Holloway (2026), the primary theoretical anchor for ProxiShop's architecture.
Channelling
AI provides voice adapted information access. Products narrated. Questions answered. Cart managed, entirely through natural speech.
Coordinating
AI mediates the workflow between the customer and the handler. Real-time guidance, session management, and intent routing.
Co-Creating
Customer, handler, and AI together complete the shopping session as a three-way bounded partnership toward a shared goal.
Academic Context
ProxiShop and the Fifth Era of E-Commerce
Building on Chu et al. (2007)'s four era framework, this dissertation proposes a Fifth Era, defined by real time human and AI co-delivery, voice-led interaction, and accessible by design as infrastructure.
Select Bibliography
Peer-reviewed foundations
Xiao, L. and Holloway, C. (2026), Channelling, coordinating, collaborating: a three layer framework for disability centred human agent collaboration. ACM CHI 2026. Primary theoretical anchor for ProxiShop's architecture.
Chen, Y. and Li, X. (2024), Expectancy violations and discontinuance behavior in live streaming commerce. Behavioural Sciences. Validates human handler model over AI only streamers.
Jindam et al. (2023), MEDSEE: a voice user interface for e-commerce for visual impairment users. SSRN. Directly informs MicCoordinator STT stability architecture.
McKinsey & Company (2023), Bridging another digital divide: accessibility for blind and low vision consumers. Confirms pervasive barriers and commercial opportunity.
Gao et al. (2024), AI or human? The effect of streamer types on consumer purchase intention in live streaming. IJHCI. Demonstrates human parasocial trust advantage.
Ferraro et al. (2024), The paradoxes of generative AI enabled customer service. Business Horizons. ProxiShop resolves the empathy and quality paradoxes through human handler model.