Generativ AI för effektivare design- och planeringsprocesser

Important results from the project

** Denna text är maskinöversatt ** Through agile working methods, we measured results and goal fulfillment on a number of test plots together with involved actors. By defining quality as "how closely the output follows the intentions of the proposal," we were able to measure and observe significant improvements in quality. The project´s technical developments, including the integration of location data, design strategies, AI, and improved "generate like" processes, have significantly improved both proposals and process. We achieved around 15% time savings in the early stages.

Expected long term effects

During the project, we have offered a free AI solution for civil engineering projects, which has anchored the use of AI in the sector. The goal was to onboard 20 municipalities, and we engaged a total of 22. The project has shown that AI can improve the planning process and confirms the interest in AI in urban planning. Out of 100 municipalities contacted, 28 planning architects from 15 municipalities have created accounts. Another 7 municipalities participated through feedback and workshops.

Approach and implementation

** Denna text är maskinöversatt ** We started with site analyzes and analysis of AI use in the planning process. We then developed design strategies and integrated OpenAI API. We also developed functions to generate similar suggestions. In parallel, workshops were conducted with architects for ongoing improvements, followed by dissemination of results. The approach contributed to the project´s objective and provided insights into how AI should function as an autocomplete rather than generating 100% ready proposals.

External links

Hemsida där vi berätta om plattformen, samt delger artiklar från projektets samarbetspartners