OpenAI says Heimstaden is using its AI tools to streamline property operations, with a Codex-assisted review of 20,112 assets and virtually staged images delivering reported efficiency gains.
In a case study published Nov. 3, OpenAI said the Codex work covered Heimstaden’s property assets. It estimated that the assessment could be done by two people over two months, compared with five people working for a year. Those figures describe the estimated effort, not a confirmed account of who completed the work. OpenAI also said Heimstaden avoided outside consultants for the assessment and that scattered data had made the task labor-intensive.
OpenAI said Heimstaden’s leaders wanted to empower employees and rethink business processes. The rollout included staff training, an AI transformation office and a network of AI champions, alongside work on property-management systems and data foundations.
In the Netherlands, producing virtually staged property images cost more than 99% less, according to the case study. The case study said Heimstaden estimates annual savings potential of €130,000 from virtually staged images in the Netherlands.
Heimstaden evaluated platforms in June 2025 and started its internal ChatGPT Enterprise rollout in September 2025. By two to three months after launch, half of its employees were using the service, OpenAI said.
The case study also describes a procurement platform that uses APIs to access OpenAI models and connects workflows across Outlook, SharePoint and Teams. According to OpenAI’s case study, Heimstaden employees make the commercial decisions and provide final sign-off.
OpenAI’s case study says Heimstaden is co-developing Amplio Estate for AI-enabled workflows using the Model Context Protocol (MCP). The protocol lets compatible AI applications connect with external tools and data. OpenAI said the work includes human oversight; Heimstaden describes more autonomous buildings as a future goal.
