AI can read the lease, but It can’t take the vendor to lunch

By Tom Entwistle, LandlordZONE

Artificial Intelligence (AI) is the new industrial revolution of our age. Some say it will dwarf in scale and impact the coming of the railways or the internet. Any activity where data on a large scale is involved – and property is very much about large data - it's meat and drink to AI.

Artificial intelligence is already speeding up administration and due diligence on commercial property deals. Now, one property lawyer suggests AI use could go even further and change the way commercial property deals are structured. From overage agreements to rents linked to live building data. Is he right?

It’s a question every commercial landlord should be asking. They should consider this: If everyone has the same machine reading the same data, on the same large language models, where will an advantage come from? 

After long experience in the industry, of buying, developing and letting property, my answer is, it will still come primarily from where it always has – from people contacts.

The argument comes from Nick Leavey, a commercial property partner at Surrey law firm Morr & Co., and he has a point: 

“Everyone is talking about AI making due diligence quicker, and that is already happening,” he says. 

“But I think the really interesting question is what happens next. If we can analyse much more information, much more quickly, it could start to change the way we structure commercial property transactions in the first place.”

It is a thoughtful point based on sound logic and deserves a hearing. It also deserves some scrutiny from the landlord’s side.

References here to the courts, the Law Society and solicitors’ duties relate to England and Wales. The RICS standard applies to RICS members and regulated firms worldwide. This article is for general information only and does not constitute legal, tax or financial advice. Landlords and property professionals should take independent professional advice for their own circumstances.

Where is AI already earning its keep?

Since ChatGPT was launched to the public just over 2 years ago, take-up of the use of AI in business has been rapid. Property software company Re-Leased reported in August that 65 per cent of the UK commercial property landlords it surveyed were now actively using AI. That’s up from 22 per cent a year earlier, though most of them (52 per cent) were still at the limited or pilot stage. 

The most common use, cited by 49 per cent of active users, was lease administration and abstraction, followed by rent collection, financial management, and portfolio reporting. Re-Leased sells the technology concerned, so its findings should be read with that in mind.

The global experience is similar. In an industry that is traditionally people focussed change takes time. Surveyors are not generally tech savvy people, they are, in my experience, primarily people focussed and that’s their key skill set.

JLL’s 2025 Global Real Estate Technology Survey of more than 1,500 senior investors and occupiers found that 88 per cent of investors, owners and landlords had started AI pilots, typically running around five at once. 

This approach lags well behind enthusiasm, however. Only 5 per cent of occupiers said they had achieved all their AI programme goals, and JLL judged more than 60 per cent of investors to be unprepared strategically, organisationally and technically.

None of that is a criticism of the people or the technology. Reading a 100-page lease to extract the break dates, rent review mechanism, repairing obligations and alienation clauses has always been slow and expensive work. 

If AI bots can do the first pass in minutes, the smaller landlord, surveyor’s practice or occupier without a back-office administration team probably stands to gain the most. I would use it, and I suspect most of our readers are already starting to do so.

Faster deals but human involvement 

Leavey’s argument is that AI could change what is agreed, not just how quickly the work is done. He gives three examples: 

1. In development agreements with overage or deferred payments. A landowner receives more if a developer secures better planning consent or achieves higher values, AI could help identify where extra value might arise and then monitor when a payment has fallen due. 
2. Turnover rents, already common in retail, could become easier to assess and administer with better business and market data. 
3. And green lease clauses. These often amount to little more than an agreement to cooperate, they could become “much more specific” once buildings produce reliable data on energy and water use.

These are sound ideas, and I can see the attraction. But each depends on something no algorithm can supply. It needs one party to be willing to share commercially sensitive information with the other. 

A turnover rent only works if the tenant opens its books. An independent retailer is not going to hand its trading figures to a landlord it doesn’t know personally. Overage monitoring depends on the developer reporting its sales honestly, and green lease targets depend on the tenant agreeing to be monitored. A data-linked lease is, at its core, a relationship of trust written into a contract.

All this may matter even more soon. Once the ban on upwards-only rent reviews takes effect, landlords are likely to look much harder at fixed uplifts, indexation and turnover-based structures. As I discussed in my recent article on the subject, performance-based rents will be negotiated between people before any AI analysis measures them.

What AI can’t know

AI works on data that already exists and has been recorded. Land Registry prices, rental comparables, planning portals and the commercial databases. But much of what actually decides a commercial deal is never written down; it comes down to experience and judgement. 

Why is the vendor selling, and how quickly do they need the money? Is the tenant’s business quietly struggling, or about to expand? Is the high street improving or declining, and which anchor tenants are likely to stay? How will the local planning authority likely view a change of use? Which agent will come up with a suitable tenant?

In my experience, many of the best opportunities in regional markets never reach the open market at all. They are offered quietly to buyers who are known to perform. Those operators who compete when they say they will, who don’t chip the price at the last minute, and who have done business with the agent or vendor before. That reputation takes years to build, and no software can earn that for you.

There is also a paradox here. As AI analysis is cheap and universal, it stops being a competitive advantage. If every bidder runs similar models over the same data, they will tend to arrive at similar factors and prices. The edge then shifts to people contacts and information that isn’t in the universal dataset. 

The professional practitioners tend to agree. A survey of 255 commercial real estate professionals by First American Data & Analytics and DealGround, carried out in the US in April 2026, found that 66 per cent used AI weekly or daily, but only 5 per cent fully trusted it enough to inform real deal decisions. 

More than half kept it to support work only. At a Savills discussion meeting in New York this summer, one broker summed it up very neatly. “What AI can’t do is make a friend.” Another broker observed that no client would hire or trust an adviser “solely because of their AI capability.” Both sources here are American, but I doubt the picture is very different in Manchester, Leeds or London.

Faster may not be safer

To his credit, Mr Leavey is clear about the risks. “Time pressure always adds risk because there is a greater chance something gets missed,” he says. “Something can be technically correct without necessarily being right for a particular client, property or transaction. That is where professional judgement is still going to be extremely important.”

The courts have already shown what happens when that judgement is missing. In R (Ayinde) v London Borough of Haringey [2025] EWHC 1383 (Admin), the Divisional Court dealt with a barrister’s grounds that cited five non-existent cases, alongside a separate claim in which 18 of 45 citations turned out not to exist. 

The court warned that freely available generative AI tools “are not capable of conducting reliable legal research” and referred the lawyers involved to their regulators. If a tool can invent case law, it can misread a break clause.

AI is potentially very capable of being very accurate but is perhaps at an early stage of its development. It is well known that if it can’t find the information it’s looking for to complete its task, it will invent it itself. Experienced human oversight is essential in professional settings. 

The professional bodies have responded to the dilemma. The RICS standard on the responsible use of AI in surveying practice became mandatory for members and regulated firms on 9 March 2026. 

Surveyors must now make clear to clients, “in writing and in advance”, when and for what purpose AI will be used. Firms must keep a written register of the AI systems that materially affect their services, and any reliance on an AI output must be signed off by a named, appropriately qualified surveyor who accepts responsibility for it. 

The Law Society’s guidance for solicitors, updated in September 2025, is equally sharp.  Using AI “does not absolve you of legal responsibility or liability” if the results are wrong, and confidential client data should not be put into free online tools.

The same caution should apply to landlords. Pasting draft heads of terms or a tenant’s accounts into a free chatbot may breach confidentiality, and where the tenant is a sole trader it may raise data protection issues too. 

Mr Leavey warns that, “The biggest mistake businesses could make is assuming AI is going to be 100% right.” In that he’s absolutely right and anyone using it should be aware of the other warnings as well.

In conclusion

AI use should be for the first pass on financials, leases, reports and comparables, but a qualified person must check anything intended to be relied upon for decision making or incorporating into legal documents. 

Professionals should check that their terms of engagement cover AI use and that their professional indemnity (PI) insurance covers for it.

AI will no doubt change the administration work of commercial property, and it may well change how some deals are structured but the caveats pointed out above cannot be ignored. Landlords, agents and other property professionals do so at their peril.

[Main Image Credit: Brett Sayles]