Conversational AI Agents in Physical Spaces: Implications for the Orchestration of Interactions & Service Experience
Executive Summary
The presentation explored the use of conversational AI agents in public physical spaces—a contrast to most current AI applications, which focus on digital, one-on-one, and often asynchronous interactions. The research, conducted in a garden centre setting, studied how customers engage with an AI agent that provides plant care advice and revealed principles for successfully integrating AI into social, shared environments.
Background & Motivation
- Most AI-human interactions happen privately (e.g., ChatGPT), not in shared physical spaces.
- Yet much of human activity is social and public, creating a gap in research and design.
- Workers also demand guidance on how to integrate AI into workflows (McKinsey survey: ~50% expect AI literacy support).
Case Study
- A touchscreen-based conversational AI deployed in a garden centre, answering plant-related questions.
- Observations and interviews with real customers provided insights into user behaviour and social dynamics.
Key Findings – Four Winning Principles
- Treat AI as an interaction partner, not a tool: Differentiate AI agents from ads or kiosks, make interactivity obvious, and communicate clear value (speed, context, stock info).
- Engage untapped users: Some people avoid human advisors out of shyness or fear of judgment; AI lowers these barriers. Peer observation helps normalize usage (“seeding strategy”).
- Foster agency and ownership: Users can adapt AI behaviour (tone, length, content), but often don’t realize this. Designing for personalization makes AI unique compared to fixed kiosks.
- Shape the cyber-social interaction: Multiple people using AI together creates challenges (who prompts, group dynamics, judgment of phrasing). AI yet lacks social feedback, but design features (e.g., visible “thinking” pauses) improve acceptance.
Business Implications
- In retail: Potential to increase(s) sales, basket size, and reduce(s) churn by offering faster, always-available advice.
- In organizations: Can amplify unheard voices, stimulate(s) team discussions, and enhance(s) cross-functional problem-solving.
- Social adoption benefits from context-specific design, personalization, and managing group dynamics.
Conclusion
Conversational AI in physical spaces has significant potential beyond retail—extending to professional teams and workplaces. To succeed, solutions must account for social norms, shared use, and human factors, not just technical performance.
