Abstract
This study investigates how agentic AI can support people with low vision (PLV) during the underexplored yet critical stage of pre-trip planning for visits to open spaces. A two-stage qualitative design—ten semi-structured interviews followed by a co-design workshop with two participants—identified key challenges: fragmented and inaccessible information, psychological burdens and dependence on others, limitations of existing assistive technologies, and expectations for proactive but controllable AI support. The findings show that travel decisions are shaped by uncertainty, fragile trust, and the need for timely, transparent, and personalised information. Participants valued companion-like interaction styles and layered control over system proactivity. The study reframes pre-trip accessibility as both functional and psychosocial, demonstrating how agentic AI can reduce cognitive load, strengthen autonomy, and enhance confidence, offering new directions for inclusive AI design.

This publication has 23 references indexed in Scilit: