|
Getting your Trinity Audio player ready...
|
This is the second article in a series about artificial intelligence and the future of real estate brokerage. The first article – Portals, CRMs and traditional agents: Which part of the real estate industry will survive the autopilot? – examined the structural change of the entire industry. This article takes a deeper look at the role of the estate agent – and what this structural change means. These are my own conclusions from building properti. The research confirms what I have observed.
I hear one question again and again at real estate conferences, from brokers, from investors, from competitors: “Will AI replace the broker?”
In my opinion, this is the wrong question. And I know it’s the wrong question because I’ve been involved in over 11,000 transactions in one of the most challenging real estate markets in Europe – which gives me a perspective that no researcher, no matter how diligent, can fully match.
The right question is: Which parts of today’s brokerage activities will AI take over better – and which parts will become more valuable precisely because AI eliminates everything else?
My answer is neither optimistic nor pessimistic. It is structural.
McKinsey estimates that the use of AI in knowledge-based work in the real estate sector can create annual added value of 430 to 550 billion US dollars worldwide. Sequoia Capital described real estate brokerage as one of the biggest unsolved autopilot opportunities worldwide. I also spoke to over 20 founders who are rebuilding their organizations around AI and found that some are planning to reduce their engineering teams from 120 to 25 people – not because the work is disappearing, but because people are moving up the value chain.
And here’s the uncomfortable truth that many leaders in brokerage organizations don’t say: The majority of today’s brokers are not consultants. They are document processors. AI is not replacing these brokers. It merely makes visible what has always been the case – that the supposed added value was never really there.
The two components of every transaction
McKinsey & Company’s agentic AI framework introduces a distinction that every real estate professional should know: the difference between (work) steps and thoughts.
“Work steps are repeatable tasks that benefit from speed, consistency and clear handovers. Thoughts are decisions that require discretion: Exceptions, trade-offs, expressions of taste or creativity, actions to build trust, and decisions involving financial, reputational or regulatory risk.” – McKinsey, March 2026
Sequoia Capital calls the same thing “Intelligence Work” and “Judgement Work”. The terms are different, the insight is identical.
Every real estate transaction consists of a combination of steps and thoughts. The steps can be automated. The thoughts are not – and even gain in value the more AI takes over the rest.
The problem is that agents today spend 70-80% of their time on steps: Creating valuations, writing listings, processing follow-ups, scheduling viewings, tracking documents, updating CRMs, coordinating administrative processes.
These are not discretionary decisions. They are repeatable, rule-based tasks that AI already performs faster, more consistently and more cost-effectively.
The consequence is clear: consistently automate steps. Protect thoughts in a targeted manner.
A broker who does not understand this distinction is not only inefficient. It is structurally obsolete.
What I have learned from over 11,000 transactions
Every transaction I have been involved in building – as a founder, as a platform operator and as someone who works with both data and people – has confirmed the same insight: The key value in real estate brokerage comes not in the execution of standardizable tasks, but in situations where judgment is required.
This is where operational processing and genuine consulting are separated. Valuations, advertisements, follow-ups, scheduling and document coordination can be standardized and increasingly automated.
What cannot be automated is the right approach at the right moment: not pushing a salesperson too early, correctly classifying a risky high bid or setting the decisive impulse in a deadlocked negotiation.
Agents who recognize this change early on and consistently focus their time on these value-creating moments are not competing in a shrinking market, but in a growing one. This is precisely the Jevons Paradox in the real estate industry: the more automation, the more valuable human judgment becomes, which cannot be automated.
The architecture that makes this change possible
Precision is crucial here, because most discussions about AI and real estate fall short.
The transition from steps to thoughts doesn’t happen by itself. An independent broker with six fragmented tools can’t eliminate the steps – he’s too busy executing them to develop the skills that make him irreplaceable.
What is required is what Benedict Evans calls the “cars on gravel roads” problem. The car existed. But its value could only unfold when roads, highways and cities were designed for it. Putting a new technology on old infrastructure only leads to incremental improvements – not transformation.
That is why I am convinced that the future of real estate brokerage is not the independent broker with AI tools. It is a two-party architecture – and only two parties – with clearly separated functions.
These two parties are the platform and the broker. Not four roles. No management level. Two.
The platform: more than AI
properti combines two central components: an intelligence layer (AI-supported – evaluation, lead scoring, content, follow-up, decision logics) and an operational engine (centralized – legal & back office, brand, lead generation, training). Together, they eliminate 100% of non-relationship-related work from the broker’s daily routine.
Without a platform: The broker spends 70% of his time on work steps. No cumulative intelligence. No scalability beyond 5-8 deals per year. No measurable quality standard.
The broker in the market
He is responsible for the relationship level: the first meeting, building trust, understanding the actual sales intention, recommendations between competing offers, negotiations in deadlocked situations and personal responsibility for the result.
Without an estate agent: no transaction. In Switzerland, anyone can call themselves a real estate agent – without a license, without an exam, without minimum standards. This makes reputation and personal responsibility the only real quality signal for owners.
Neither party can replace the function of the other. The platform cannot build trust in the initial meeting. The broker cannot accumulate decision-making intelligence from over 11,000 transactions. Together they create something that neither a single broker nor a pure software solution can achieve.
This corresponds to what Foundation Capital describes as a “context graph”: a searchable structure that records how and why decisions were made – not just what happened, but under what conditions, with what strategy and with what result. Every transaction at properti expands this system. Every broker benefits from all previous transactions.
What “thoughts” look like in practice
Here are some concrete examples to illustrate this:
The saleswoman who is not yet ready
A 67-year-old woman has lived in her house for 34 years. Her children are urging her to sell. She has agreed to advertise the property. But in the first conversation, she doesn’t ask about the price, but about the garden, the neighbor’s tree, the light in the kitchen in the morning.
The platform rates her as a highly likely seller. The property is optimally positioned. The documents are prepared.
A good estate agent recognizes that she is not yet ready. An immediate closing leads to a failed sales process. The right decision is to wait and see.
The offer that is deceptive
Three offers have been received. The highest is 5% above the price.
An experienced estate agent checks the financing. The buyer is at the limit. A bank valuation could be lower. The risk of failure is high.
A lower supply offers more security. The platform provides data. The broker makes the decision.
The negotiation that requires presence
The final offer is lower than expected. The seller wants to re-advertise.
The broker has 15 minutes to understand the situation: Is it a boundary or a negotiating tactic?
This assessment is based on experience. The implementation is based on trust.
A market that is already splitting up
This is not a forecast for the future. It is already a reality.
On the one hand: brokers who have recognized the change. They work on platforms, use data and focus on advice, negotiation and trust. They achieve 35, 50 or more deals per year.
On the other side: brokers who continue to act as information brokers. They spend 70% of their time on tasks that AI does better.
In Switzerland, more than 5,000 brokers achieve fewer than five transactions per year. A large proportion of this activity will disappear – not because AI is replacing brokers, but because those who adapt are crowding out the others.
The key question for brokers today
Not: “Will AI replace me?”
But: “How much of my time do I spend on tasks that AI already does better – and what would I do with that time?”
If this proportion is over 50%, the platform question is not optional. It is urgent.
The transformation from information broker to transaction strategist does not happen by itself. It requires an infrastructure that takes over all the steps.
I am convinced that such an infrastructure – truly AI-first – not only changes the number of transactions, but also the role of the broker themselves. More present. More decisive. More valuable.
This is the model that properti is developing in Switzerland and scaling up in Europe.
That is what properti is counting on.
Brokers who find this platform and embrace change will not only survive the next five years.
They will shape them.