en
Buyers and sellers are now using AI to build their property offers and arguments. Fairway examines how the role of the real estate agent is evolving towards analysis, arbitration and negotiation.
A new phenomenon is emerging in real estate negotiations: buyers are using artificial intelligence to draft their offers. Sellers are responding with artificial intelligence. Each side arrives with structured arguments, comparable properties, financial reasoning, legal considerations and market analysis.
On paper, this looks like progress. Exchanges are better written. Objections are more structured. Counter-offers appear more rational.
In practice, the effect is more ambiguous.
Negotiations are not necessarily becoming more efficient. They can even become longer, more theoretical and more defensive. Everyone now has ten arguments to justify their position. The buyer asks AI why the price is too high. The seller asks AI why the price is justified. Both answers can be well written, convincing and yet completely incompatible.
The issue is therefore no longer simply access to information. The issue is becoming the hierarchy of information.
And this is precisely where the role of the real estate agent is changing.
The use of generative artificial intelligence has rapidly become part of everyday practices. In France, the 2026 Digital Barometer indicates that 48% of French people already use generative AI, with particularly strong adoption among young adults, executives and self-employed professionals. The same survey indicates that ChatGPT is the most widely used generative AI service among French users.
This is not insignificant for real estate.
Buying or selling an apartment involves exactly the uses for which AI has become popular: understanding a situation, preparing an argument, drafting a message, analysing data and comparing different options. A study conducted by OpenAI with researchers from Harvard and Duke found that nearly 80% of ChatGPT usage falls into three broad categories: practical advice, information seeking and writing.
A property offer is precisely that: information, argumentation and writing.
A buyer can ask:
“Draft an offer of €1,250,000 for an apartment listed at €1,320,000, explaining that the price is too high given the DPE rating, renovation work, the market and rising interest rates.”
The seller can respond:
“Draft a counter-offer of €1,290,000, explaining that the property is rare, well located, suitable for a family, difficult to compare with average sales in the area and that demand remains strong.”
Both texts will probably be good. That is precisely the point.
For a long time, a low property offer might arrive with a fairly simple justification:
“We really like the apartment, but our budget is limited.”
Or:
“Given the renovation work required, we are offering this price.”
Today, an offer can look more like a negotiation memorandum.
The buyer may mention the price per square metre, the DPE rating, estimated renovation costs, the condominium, service charges, floor level, absence of a lift, exposure, DVF sales data, competing listings, observed marketing periods, borrowing costs and resale risk, sometimes even applying a theoretical discount by default.
The seller, for their part, can produce an equally detailed response: quality of the address, rarity of the layout, high floor, prestigious building, period character, lack of overlooking neighbours, school catchment area, heritage value, local market conditions, scarcity of large family apartments and replacement cost for purchasing an equivalent property.
As a result, the negotiation can sometimes become a battle of files.
But in a real estate transaction, it is not the party producing the most arguments that signs. It is the party that succeeds in distinguishing the arguments that genuinely affect the price from those that merely serve to defend a position.
The main danger is not that AI writes badly. The danger is that it can write an already established position too convincingly.
A buyer convinced that a property is overpriced can ask AI to build the strongest possible argument against the asking price. A seller convinced that the property is worth its asking price can request the opposite argument. Each receives an answer that reinforces their initial interpretation.
This is the classic confirmation bias: people search for, select and value information that supports what they already believe. Recent research into generative chatbots highlights that these tools can reproduce or amplify this bias when used to defend a pre-existing position.
In real estate, this bias is particularly powerful because a transaction is never entirely rational.
The buyer is afraid of overpaying.
The seller is afraid of selling too cheaply.
The buyer looks at the budget.
The seller looks at their attachment to the property.
The buyer looks at the defects.
The seller looks at the potential.
The buyer reasons in terms of total cost.
The seller often reasons in terms of the expected net selling price.
AI can help formulate all of this. It does not automatically resolve the disagreement.
In a real estate negotiation, deadlocks rarely come from a single figure.
They result from a combination of price, timing, financing, ego, perception of risk and trust.
A buyer may make an offer of €1,250,000 on a property listed at €1,300,000. Their argument may be strong: renovation work required, an imperfect DPE rating, high service charges, a more selective market and more expensive financing.
But behind this offer, the agent needs to understand something else:
The same applies to the seller.
A seller may reject an offer by explaining that their apartment is rare. That may be entirely accurate. But the agent needs to look at the facts:
AI can produce an answer. The agent must read the situation.
The wrong response would be for the real estate agent to enter into a competition of arguments with clients.
A buyer arrives with a four-page memorandum?
The seller responds with five pages?
The agent produces six pages?
That would be a mistake.
The agent's value is not to add noise to noise. Their value is to reduce the noise.
Their role is to bring the discussion back to a few very concrete questions:
This is a function of arbitration.
The agent does not replace AI. The agent puts AI back in its proper place: a formulation tool, not a mediator.
In an AI-assisted real estate negotiation, the agent needs to classify arguments into three categories.
1. Facts that genuinely affect the price
Certain factors have a real impact.
These issues need to be considered objectively. They may justify a negotiation, but their actual weight still needs to be measured.
An E rating does not have the same impact on a dark ground-floor apartment requiring renovation as it does on a beautiful family Haussmannian apartment with a lift, perfect layout, sought-after address and scarce comparable supply.
The defect exists. The question is: how much is that defect worth in this market, for this property, at this address, at this particular time?
2. Arguments that appear rational but are poorly used
This is probably the area where the agent becomes most useful.
A classic example is the average price per square metre.
A buyer may say: “The neighbourhood is at €11,500 per square metre, so your apartment cannot be worth €13,000 per square metre.”
This argument may be relevant. It may also be very weak if the average includes ground-floor properties, apartments requiring renovation, small units, buildings without prestige features, less desirable streets and older sales.
Another example is DVF sales data.
Public data is useful. But it does not always reveal the condition of the property, the quality of the layout, natural light, floor level, presence of an outdoor space, renovation standard, competitive pressure at the time of sale or specific transaction conditions.
The agent must therefore avoid two extremes:
3. Emotional arguments disguised as analysis
This is often the heart of the matter.
A buyer may say: “The property is too expensive.”
They may actually mean: “I am afraid of making a mistake.”
A seller may say: “I am not in a hurry.”
They may actually mean: “I need reassurance about the quality of the buyer's offer.”
A buyer may multiply technical reservations.
They may simply be seeking a margin of safety.
A seller may refuse even a small reduction. They may simply want to avoid feeling that they gave in too quickly.
AI does not always see this. The agent hears it during viewings, on the phone, in silences, in hesitations and in the way each party talks about their project.
The more arguments the parties produce, the more important it becomes to have someone who can return to reality.
Reality is the number of viewings.
Reality is the quality of buyer files.
Reality is financing.
Reality is the time required to sell.
Reality is active competition.
Reality is what visitors repeatedly say after each viewing.
Reality is the price at which a solvent buyer is prepared to sign now.
This is why digitalisation does not mechanically eliminate the role of the intermediary. In a highly digitalised market such as the United States, the National Association of Realtors' 2025 report still indicates that 88% of buyers purchased through an agent or broker, with buyers particularly seeking help finding the right property, negotiating terms and managing the transaction process.
The US market cannot be directly transposed to France, but it demonstrates that mass access to information is not enough to eliminate the need for intermediation in a complex transaction.
The real estate agent therefore does not derive their value solely from the information they possess. Their value lies in their ability to interpret, weigh and turn information into actionable decisions.
AI can help prepare an offer.
It can help draft a counter-offer.
It can summarise data.
It can formulate arguments.
But it does not actually visit the property.
It does not sense a buyer's level of interest.
It does not know whether silence on the phone is hesitation or strategy.
It does not know whether the seller is genuinely prepared to wait three months.
It does not know whether a buyer who criticises the property extensively is actually the person who likes it most.
It does not know whether a slightly lower offer without a financing condition is better than a higher but fragile offer.
It does not always know how to distinguish a useful comparable from a misleading one.
Above all, it does not bear responsibility for the sale.
The agent does. The agent must deliver a result: bring two parties together, secure an agreement, avoid an endless negotiation and support the transaction through to signing.
This is probably the most important point.
A real estate negotiation is not a contest of arguments. It is not a trial. It is not an essay.
The goal is not for the buyer to prove that the seller is wrong.
The goal is not for the seller to prove that the buyer is unreasonable.
The goal is to find a price and terms that both parties can accept and sign.
This is why the agent sometimes needs to bring certain discussions to an end.
But at some point, arguments need to become a decision.
This is where the real estate agent becomes central again.
When faced with a highly detailed offer, the agent should avoid responding emotionally.
The right method is to review the points one by one, but not in order to challenge everything.
Step 1: isolate the real issues
The agent should identify the two or three points that genuinely justify the price gap.
For example:
The rest may be secondary.
Step 2: remove irrelevant comparables
Not all comparable properties are genuinely comparable.
An apartment sold six months earlier on a neighbouring street is not necessarily comparable if the floor, condition, layout, natural light, building or surface area differ significantly.
The agent must explain why certain examples should be retained and why others should be excluded.
Step 3: quantify what can be quantified
Renovation work should be estimated.
Service charges should be compared.
The cost of financing should be understood.
Any potential discount should be linked to observable factors.
An unquantified argument remains an impression. A quantified argument becomes a negotiation tool.
Step 4: rephrase without humiliating
A negotiation often fails because one party feels attacked.
The buyer should not give the impression that the seller has badly valued their property.
The seller should not give the impression that the buyer is opportunistic.
The agent must translate.
Instead of:
“Your price is too high.”
Say:
“The offer takes into account the genuine level of buyer interest in the property, as well as the renovation budget and the overall acquisition cost.”
Instead of:
“Your offer is too low.”
Say:
“The seller understands the points raised, but considers that the quality of the address, the layout and the rarity of the property justify a more measured effort.”
The form matters. In real estate, the form can save a negotiation.
Step 5: bring everyone back to the timetable
AI can produce arguments indefinitely. A sale, however, has to move forward.
The agent must therefore put the timetable back at the centre:
A negotiation without a timetable becomes an exchange of opinions. A negotiation with a timetable becomes a process.
AI makes clients better informed. It can also make them more confident in their own interpretation.
This is not necessarily negative. A better-prepared buyer can ask better questions. A better-informed seller can better understand the market. A better-written offer can make an exchange smoother.
But when AI is used to justify a position instead of informing a decision, it can block the negotiation.
The role of the real estate agent is therefore not reduced. It becomes more demanding.
It is no longer enough to open a door, pass on an offer and wait for a response.
The agent must now:
This is a human, commercial, psychological and technical skill.
A seller should not reject an offer simply because it is highly detailed.
A long offer is not necessarily aggressive. It may simply reflect a need for reassurance. In a market where buyers consider the total cost of their acquisition — price, fees, renovation work, financing, service charges, taxation and condominium costs — it is normal for them to seek to justify their position.
The right question is not:
“Are all their arguments correct?”
The right question is:
“Does this offer come from a serious, financeable and motivated buyer, capable of signing quickly and completing the transaction?”
A buyer who negotiates can be an excellent buyer.
A buyer who does not negotiate may disappear after three days of reflection.
A highly critical buyer may simply be the person who most needs reassurance.
A silent buyer may never make an offer.
The agent must help the seller avoid confusing the style of the offer with the quality of the buyer's file.
AI can help structure an offer. But an offer that is too long, too cold or too aggressive can produce the opposite effect to the one intended.
A seller is not an Excel spreadsheet. They have often lived in the property, financed renovations, raised a family there and built an emotional connection with their apartment or house.
A good offer should therefore be reasoned, but also humanly acceptable.
It should say:
AI can draft the text. But the buyer must remain credible, measured and respectful.
In practice, an overly demonstrative offer can create concern. It may give the impression that the buyer is already looking for conflict before the preliminary agreement has even been signed.
The real estate agent is no longer simply a listing distributor. They are no longer merely a viewing organiser. They are no longer simply a transmitter of offers.
In modern negotiations, the agent becomes an operational trusted third party.
Their role is to say:
This is an uncomfortable position because it sometimes requires challenging one's own client.
Telling a seller:
“Your property is beautiful, but the market does not confirm your asking price.”
Telling a buyer:
“Your argument is well structured, but some of the comparables are not relevant.”
Telling both parties:
“You are close to an agreement, but you are about to lose the sale because of your respective positions.”
This is where the non-AI agent retains their full value.
Not against artificial intelligence.
But above the noise it can generate.
AI provides arguments.
The real estate agent gives those arguments weight.
AI can draft an offer.
The agent knows whether that offer can succeed.
AI can compare prices.
The agent knows which properties are genuinely comparable.
AI can formulate a counter-offer.
The agent knows how to present it without closing the discussion.
AI can prolong the debate.
The agent must bring the negotiation back to a decision.
In a real estate market that is better informed, more demanding and more heavily argued, the value of the agent does not disappear. It shifts towards analysis, methodology, trust and the ability to bring both parties to signature.
Can ChatGPT be used to draft a property offer?
Yes, but with caution. ChatGPT can help structure an offer, clarify arguments and avoid awkward wording. However, the offer must remain adapted to the property, the seller, the market context and the strength of the buyer's file. An overly generic or aggressive offer can weaken the negotiation.
Does an AI-written property offer have less value?
No. What matters is not the tool used to write the offer. What matters is the coherence of the proposed price, the quality of the financing, the seriousness of the buyer, the timetable and the terms of the offer.
Can AI accurately estimate the price of an apartment?
AI can help analyse data, but it does not replace a local real estate analysis. The price of an apartment depends on many factors that raw data does not always capture: floor level, natural light, layout, actual condition, condominium, precise address, rarity, active competition, quality of demand and perception during viewings.
Why are real estate negotiations becoming longer?
Because the parties have access to more information, but not necessarily a better method for prioritising it. Buyers can produce numerous arguments to negotiate, while sellers can produce just as many arguments to resist. Without a third party capable of sorting, quantifying and reframing them, the discussion can become unnecessarily long.
What is the role of the real estate agent in the age of AI?
The role of the real estate agent is to bring the negotiation back to reality: viewings, offers, relevant comparables, financing quality, genuinely value-reducing defects, timing and signing conditions. The objective is not to produce more arguments than AI, but to select the arguments that genuinely matter.
Will AI replace real estate agents?
AI will probably replace certain tasks: drafting initial messages, summarising documents, preparing arguments and analysing simple data. But it does not replace the ability to read a human situation, negotiate, secure a transaction, manage tensions and support two parties through to signing. The more clients use AI, the more visible the need becomes for a third party capable of prioritising arguments.
Key points to remember
"AI can provide more arguments, but the real estate agent gives those arguments context, weight and direction. In increasingly informed negotiations, human judgement remains essential to turn information into an agreement."
 | Author of the article Hugues de Poulpiquet Fairway Luxury Real Estate · Real Estate Market Trained as a lawyer and founder of Fairway Luxury Real Estate, Hugues de Poulpiquet writes Fairway's market insights dedicated to the Paris real estate market, sales strategies and the valuation of high-end residential properties. |
Fairway Luxury Real Estate
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.