The customer journey is beginning to shift as AI develops, new Inman contributor Roland Kampmeyer writes, even if we cannot yet fully assess the direction or speed of the changes.
The way people begin their real estate journey is shifting dramatically. Not tomorrow, but right now.
When someone considers selling a property today, their research doesn’t necessarily start with a call to an agent or a search on a listing portal, but with a question to ChatGPT, Gemini or a search on Perplexity. “Should I sell my house now?” or “Which agent is right for my property in my neighborhood?” — these are the moments where leads are generated or lost.
In the future, AI agents will handle this search directly in AI browsers like Atlas or Comet, independently conducting initial inquiries or comparisons.
This shift affects the entire funnel. At the very beginning of the customer journey, the decision is made about who even makes the shortlist. If a potential seller seeks initial guidance from an LLM and our brokerage isn’t present there, we are out before the game even begins.
The same applies to the buyer side. If someone searches for homes in specific locations and the answer comes from an AI that doesn’t know our listings or our market expertise, we will no longer reach that person.
It’s a different story at the end of the customer journey. Valuation, personal consultation, negotiation, the entire transaction with all its legal and emotional complexities — this is where humans remain irreplaceable, at least for now. Selling a property isn’t an everyday decision like buying a plane ticket or the next pair of sneakers. Too many variables, too much regulation, too much emotion. But if we lose the top of the funnel, we never even get the chance to leverage our strengths at the bottom of the funnel.
This challenge affects everyone who wants to bring properties to clients or win sellers as clients. But the playing field is completely uneven. Depending on the listing portal, significantly more resources are available for AI functions on their platforms. Brokerages will hardly be able to provide the same technological answers; portals scale at entirely different dimensions.
At the same time, we are all facing the same black box. Today, nobody really knows by what criteria ChatGPT or other LLMs select brokers and/or properties for their answers. There are no transparent ranking factors like in classic search engine optimization, no comprehensible algorithms, no clear optimization strategies.
Anyone selling “LLM optimization” today is working with speculation and assumptions, not with reliable data. Furthermore, different LLMs use different data sources, and the mechanisms are constantly changing.
The pace of change is so high that traditional strategy cycles no longer work. This is an extremely uncomfortable position for companies that need to plan strategically.
But there is another uncomfortable question that is asked less often: Does AI-based real estate search even solve a relevant customer problem? Or is it a technical solution in search of a problem?
If someone is looking for a house in a specific location, is an AI-powered conversation really superior to the precise filter search on a good brokerage or portal website, complete with maps, pictures and direct property comparisons? And when someone is considering selling, do they want to discuss their decision with an LLM, or do they need a conversation with someone who knows the local market from years of experience and understands their specific situation?
Perhaps the real value of AI lies in completely different areas than the initial search or orientation. Or perhaps not. As of today, we simply don’t know.
What we do know is this: The customer journey is beginning to shift, even if we cannot yet fully assess the direction or speed. Doing nothing is not an option, but blindly chasing every technological trend isn’t either.
My approach, therefore, is not a master plan, but a manifesto of what I am actively doing:
My AI manifesto
- Experiment and use AI where possible and sensible — for content, marketing, data, processes and automation.
- Use the pro version of everything. Data security is the top priority. No compromises on company and client data.
- Don’t do what’s cool, do what’s useful for you and your clients. Technological tinkering leads nowhere.
- Bring your team along. AI belongs in your company’s DNA, not just in the hands of a few enthusiasts.
- Accept that no one really knows the “AI rules.” Not even the self-proclaimed experts.
- The pace of change is faster than your plan. Get used to being constantly overtaken. Act anyway.
- Don’t think and build everything yourself. Use proven AI solutions. Look strategically for tools you can integrate quickly into your company with little effort.
- Take wrong turns to find the right path. Failure is part of the process, not the end.
- Communicate openly that you are experimenting. Authenticity beats faked AI expertise.
- Treat your data and market data as a strategic asset. This is your differentiator against other providers and portals.
This isn’t a prescription. It’s my way of dealing with these fundamental new possibilities. I experiment, I fail, I learn.
I invest in pro tools because data security is non-negotiable. I try to bring my team along because AI isn’t a one-person show. And I accept that I don’t have all the answers.
Roland Kampmeyer founded KAMPMEYER Immobilien in 1995 and has since built one of his region’s leading residential brokerages with offices in Cologne, Bonn and Düsseldorf, Germany. Connect with him on LinkedIn.