The buildings haven’t changed. Everything around them has.
A hundred years ago, managing a residential building meant finding tenants, collecting rent and keeping properties in good condition. Today, these fundamentals remain remarkably similar. What is changing is how that work gets done.
For European residential real estate operator Heimstaden(opens in a new window), this presents an opportunity: what if AI could fundamentally change how thousands of homes are managed?
In June 2025, Heimstaden’s leadership decided to accelerate its bet on AI. After evaluating several platforms, the company chose OpenAI. The goal was twofold; empower employees and rethink the processes at the heart of its business.
The ambition is to move towards increasingly self-operating buildings, where AI can help anticipate needs, coordinate work and support decisions across the property lifecycle. “The purpose of our buildings hasn’t changed.” says Helge Krogsbøl, Co-CEO at Heimstaden. “We provide homes for people. But the way we operate them continues to change as we challenge conventions and adopt new technology.”
“We see AI as one of the biggest opportunities to accelerate that transformation.”
—Helge Krogsbøl, Co-CEO, Heimstaden
Reimagining the business of letting homes
Consider what happens when a property becomes vacant.
Traditionally, getting it back onto the market involves assessing rental value, arranging photography, producing advertisements and responding to market demand. Each stage requires time, coordination and specialist input. And it is an information challenge; understanding what you own, what the competition has, and what the customer wants.
AI is now supporting all the key steps in this process:
Pricing: Intelligence and recommendations based on tools built in ChatGPT Sites
Advertising: Image styling and apartment descriptions done by AI
Review listings: Agent reviews all live ads against Heimstaden’s quality standards
Feedback: AI generated market feedback helps letting teams take corrective action which makes prices increasingly dynamic
Employees remain responsible for reviewing and approving decisions. But the work surrounding those decisions is changing.
In the Netherlands, the cost of producing a virtually staged property image has fallen from approximately €50 to less than €0.10 which is a reduction of more than 99%. Heimstaden estimates annual savings potential of €130,000 in that market alone.
The significance extends beyond cheaper advertisements. By applying AI across the letting process, Heimstaden is building a more responsive way to bring homes to market.
“Our AI strategy is based on three pillars.” says Nikolaj Brammer, Heimstaden’s Chief Digital Officer, “First, we empower employees through training and access to the right tools. Second, we have a central transformation office and a network of AI champions to drive larger AI initiatives. Third, we strengthen the foundations by making our property management systems and data models ready for AI.”
Letting puts all three pillars into practice: tools and training enable colleagues to create virtually staged images and listings; our central team coordinates improvements across pricing, advertising, ad review and market feedback; and we are making property data accessible to the AI tools that support pricing and listing creation.
“We started by breaking down our most important business processes and asking where AI could create value. Letting is a good example: it’s not about automating one task, but improving how the whole process works.”
—Nikolaj Brammer, Chief Digital Officer, Heimstaden
When domain experts become builders
Some of the most significant changes are happening far from traditional software engineering teams.
Heimstaden needed to assess environmental risks across its portfolio. The information existed, but was scattered across public sources, formats and datasets. Turning it into a usable assessment would have required substantial manual work and technical resources.
Heimstaden’s Julia Erlandsen, an environmental sustainability analyst without a software development background, used Codex to help build a different approach.
Codex supported the discovery, cleaning and standardisation of data, quality checking and the development of repeatable processing pipelines. Julia retained control over the methodology and validation.
The result: 20,112 assets assessed, more than six million rows of useful data generated, and an estimated delivery requirement reduced from five people working for a year to two people working for two months.
Heimstaden also avoided engaging external consultants which would otherwise have been needed to solve the problem.
What began as a complex analytical project is becoming a repeatable capability, supporting climate-risk assessments, reporting requirements and investor decision-making.
“Codex helped us build a solution that would previously have required substantial technical support. We now own the methodology and can continue developing it ourselves.”
—Julia Erlandsen, Environmental Sustainability Analyst, Heimstaden
Responsible AI adoption
Heimstaden’s AI adoption focuses on tasks employees need help with. After ChatGPT Enterprise launched in September 2025, half of employees were using it within two to three months. A central AI transformation office maintains a library of use cases so colleagues can learn from examples across the business.
Employees use approved tools within rules for data security. They remain responsible for reviewing AI-generated content before sending it and for approving decisions.
Adoption extends to real estate caretakers, including some who were initially sceptical. In Norway, one caretaker used ChatGPT to assess a neglected apartment and draft a prioritised refurbishment plan, turning observations into a structured plan they could review and refine.
“People across the business are finding practical ways to use AI in their daily work. Our job is to scale what works while keeping data secure, responsibilities clear and people in control.”
—Andreas Kalmar Lindquist, Manager AI First Team Heimstaden
The next frontier: From answering questions to taking action
Heimstaden is already preparing for the next stage: AI that can work across systems and coordinate multi-step processes.
Its Procurement Intelligence Platform, currently being rolled out, combines workflow automation with OpenAI models accessed through APIs. Designed to operate across Outlook, SharePoint and Teams, it supports contract analysis, negotiation preparation and contract lifecycle management, with employees retaining responsibility for commercial decisions and approvals.
The longer-term ambition is more fundamental.
Heimstaden is co-developing Amplio Estate, a modern property management system designed to support AI-powered workflows through MCP. The goal is to enable employees to use AI, such as ChatGPT, to retrieve property information and take action in core business systems, with appropriate human oversight.
This is how Heimstaden begins turning the idea of self-operating buildings into a practical technology roadmap.
Results at a glance
More than 99% lower image production costs in the Netherlands.
20,112 property assets assessed with Codex, reducing the estimated delivery requirement from five people over one year to two people over two months.
50% employee adoption within 2–3 months of rollout.
Leadership lessons:
Focus on people: Adoption and imagination are the bottlenecks (not tech)
Find your champions: The biggest returns come from colleagues driving change locally
Get the process right: AI will not fix a broken process, but it can dramatically scale a good one.
What’s next
Heimstaden’s vision is not of buildings without people, but of properties supported by increasingly intelligent systems which will improve decision making for employees and increase ability to resolve matters for customers.
The buildings will remain much as they have always been—homes and roofs over people’s heads. The way they’re operated has forever changed.
Instagram Engineering
Shopify Engineering