Participatory urban planning is undergoing significant changes because of the rapid development of artificial intelligence (AI) tools. This study explores and compares the differences in the relational agency of Analytical, Generative and Humanised AI as perceived by urban planners. We consider relational agency with respect to the intended and unintended uses and perceived opportunities and risks of these technologies in participatory urban planning. A Delphi survey was conducted with 34 urban planners across municipalities in Finland with > 20,000 inhabitants, followed up by a focus group and survey to build convergence on respondents’ views (n = 13). Respondents commonly used Analytical AI to summarise plan information and to report on environmental variables, and saw its potential to analyse and report on citizen feedback on draft plans, including survey responses. Generative AI is currently used to undertake more creative tasks than Analytical AI, including generating images and simulations, with the potential to use it to draft plan content and summarise expert reports on plan drafts. Risks centre around data protection problems, the outsourcing of decision-making to computer tools, and the potential to create misinformation or incorrect content. Humanised AI is rarely used in urban planning in Finland, but respondents saw potential to use it to respond to community feedback on urban plans and to communicate plans more effectively to the wider public through a cautious approach that addresses the potential for emotion and opinion influencing. We discuss the implications of these findings for participatory urban planning across Europe and elsewhere.