AI that gives real estate agents more time for clients

Yes, provided you are transparent and the text is accurate. NVM and VBO do not ban AI-generated text, but the agent remains liable for the content of the advert. An AI description that mentions non-existent features or makes the property look better than it is, falls under misleading advertising. Our approach: AI drafts a concept based on your own object data (size, year, features from the CRM), the agent reviews and adjusts. For AI-edited photos (virtual staging, tidied gardens) it gets stricter: NVM and consumer organisations push for clear labelling of edited images, and the AI Act in some cases requires disclosure that the content is manipulated.

Yes. Most Dutch real estate systems offer integration via NVM exchange, MAPI or their own API. We build or configure AI tools that pull object data from your CRM, generate a Funda description in your office style, and write the draft back to the CRM for your final check. The same applies to lead data: a chat intake on your site can ask qualifying questions, push a score into your CRM, and route to the right agent. We work with your existing system, not against it. For lesser-known packages we first verify that a workable integration is possible.

Buyer leads contain name, phone, search criteria, sometimes income or financing indication. That is privacy-sensitive data and falls under GDPR. We build flows where data stays within the EU, is not used for training general AI models, and is only accessible to your office. For chatbots and mail flows we document what is stored, for how long, and when it is deleted. For Wwft checks (sellers and buyers above the threshold) we work with validated Dutch providers that perform source checks within the correct legal framework. No US data parties for identity data.

No, and that is not the goal. The value of an agent is the personal contact, reading a buyer, negotiating, and knowing the neighbourhood. AI takes over the work around it: the Funda text, the confirmation emails, writing up the viewing report, the follow-up after an open house, the bidding status updates to sellers. That gives you more time for the work clients actually come to you for. Same for lead qualification: a chatbot handles the first questions, you only call when someone has serious interest and a clear search profile.

That is a real risk, so we design around it. First: the AI only writes from object data you have entered into the CRM. It does not invent features that are not in the source. Two: every draft goes through human review before it lands on Funda. Three: for figures with legal impact (area according to NEN 2580, year built, energy label) we use the values from the measurement report, not a model estimate. Four: we log which version of which model produced which text, in case questions arise about an advert later. Liability stays with the office, so the control loop must hold.

That is exactly where the time saving is biggest. A solo agent or small office does not have a junior or marketing person to handle Funda text and mail flows. With AI you do that yourself in a fraction of the time. A description for a typical family home that you would normally type up in the evening is ready for review in minutes. Viewing confirmations and follow-ups run automatically instead of from memory. You can start small: a solid Funda text generator linked to your CRM, a WhatsApp flow for leads outside office hours, and only later expand to lead scoring or market analysis.

We start with a short analysis at your office. We look at where the most time is lost: is it the Funda text, the evening follow-up on leads, writing up viewing reports, or for instance the zoning plan extracts for commercial property. Then we build a first tool around the biggest pain point and connect it to your CRM. You test it in practice, give feedback, we refine the prompts and tone until it really fits. Only then do we expand. No big one-off implementation, but step by step so you stay in control and the office adapts to the new way of working.