A large acquisition can carry thousands of pages of leases, engineering reports and financial statements, and a deal team has weeks to read all of it before the money moves. Miss one clause in one lease and the return model can be wrong.
That pressure, high stakes against a hard clock, is exactly where AI has started to change how institutional investors work. Due diligence has become a test of how well a firm pairs its analysts with tools that read faster than any person can.
This article covers what due diligence involves and how AI is compressing the timeline. It then looks at where AI is applied, across document review, physical assessment, market analysis and risk, and the limit where human judgement still decides.
What Does Due Diligence Mean in Real Estate? (A Refresher)
Due diligence is the investigation an investor runs on a property before committing capital, to confirm the asset is worth what the deal assumes and to surface anything that could change the price or kill the deal.
It spans four main workstreams:
- Financial: checks the income and expenses, the rent roll and the leases, to confirm the numbers the deal relies on.
- Physical: assesses the building's condition and the cost of any repairs or capital works.
- Legal: reviews title, zoning, contracts and anything that could cloud ownership or use.
- Market: tests whether the assumptions about rent growth and demand hold up.
Each workstream feeds the same decision, whether to proceed, at what price and on what terms. The work is detailed, deadline-bound and expensive, because a team of well-paid professionals reads a mountain of documents under time pressure. This is why the parts AI can accelerate matter so much to the economics of a deal.
How AI Is Compressing the Due Diligence Timeline
The clearest value of AI in due diligence is speed. Work that once set the pace of a deal now moves faster, which changes both the timeline and the number of deals a team can look at.
Real estate transactions generate large, fragmented sets of documents, and reviewing them has always depended on analyst hours. AI reads and structures those documents in a fraction of the time, extracting the key terms, flagging inconsistencies and organising everything into a usable form. The analyst who once spent days assembling information now spends that time evaluating it.
JLL reports that firms using AI in real estate decision-making see 23% faster transaction times and 18% more accurate valuations than those using conventional methods.
The gain compounds across a pipeline. A team that clears the document work faster can screen more opportunities and walk away from weak ones sooner, before spending heavily on deals that will not close. The economics follow from that. Faster screening means lower diligence cost per deal and more deals evaluated per analyst, which is why the largest investors have moved on this first.
AI for Document Review: Lease Abstractions, Rent Rolls, and Cap Tables
Document review is where AI has the firmest hold, because the work is high-volume, rule-bound and repetitive, exactly what machines handle well.
- Lease abstraction: pulling the key terms out of a lease, such as rent, escalations, renewal options, break clauses and responsibilities. AI extracts these across a whole portfolio of leases in the time an analyst would spend on a handful.
- Rent roll analysis: checking the rent roll against the underlying leases to confirm the income assumptions, catching mismatches between what the roll claims and what the leases actually say.
- Financial statement review: reading operating statements and reconciling them against the rent roll and leases, surfacing the discrepancies a rushed manual review can miss.
The value is consistency as much as speed. A tired analyst on the fortieth lease of a long day misses things a model reads the same way every time. The model does not replace the analyst's judgement on what the terms mean, but it makes sure nothing was skipped in the first pass.
The discipline is verification. AI extraction can misread an unusual clause, so the strongest teams treat the output as a fast, thorough first draft that a person checks, not as a final answer.
AI for Physical Due Diligence: Condition Assessment and Cost Modelling
Physical due diligence has been slower to automate, because a building has to be inspected in person. AI is changing the analysis around the inspection rather than replacing the inspector.
On condition assessment, AI can read inspection photos and drone or satellite imagery to flag issues at scale, from roof wear to facade problems, giving a first read before anyone walks the site. JLL has built models that extract building quality signals, including facade quality from street-level imagery, a task that was once slow and expensive to do by hand.
On cost modelling, AI helps estimate the cost of repairs and capital works by drawing on large datasets of comparable projects. This turns a rough manual estimate into a data-grounded range, sharpening the numbers that flow into the underwriting.
The inspector still walks the building and signs off. What changes is that they arrive informed, with the obvious issues already flagged, so their time goes to the judgement calls a photo cannot settle.
AI for Market and Comparable Analysis in Acquisitions
An acquisition is only as sound as its assumptions about the market, and this is where diligence connects to the wider question of where a market is heading.
AI runs comparable analysis at a scale manual work cannot match, pulling recent transactions and adjusting for differences to test whether the deal price sits right against the market. It reads demand signals, from employment and migration to supply pipelines, to check whether the rent-growth assumptions in the model are credible.
The point is stress-testing. A deal team can use AI to challenge its own assumptions quickly, asking whether the underwriting holds if demand softens or supply builds. That check used to take a research team days. It now runs in a fraction of the time, so it happens on more deals and earlier in the process.
AI for Risk Scoring: Flagging Issues Before They Cost You
The highest-value use of AI in diligence is catching the problem that would otherwise surface after the deal closes. AI risk scoring reads across all the diligence inputs and flags the issues that need a human's attention.
A model can surface a lease with an unusual termination right, a tenant whose payment history signals trouble, an environmental flag in an old report, or a market showing signs of a turn. None of these is a verdict. Each is a prompt for the deal team to look closer at something they might otherwise have reached too late.
The value is timing. Finding a deal-breaker early saves the diligence spend on a deal that was never going to work, and finding a hidden risk before closing is the difference between pricing it in and inheriting it.
This is where AI earns its place, not by making the call, but by making sure the team sees what matters while there is still time to act.
What the Largest US Institutional Investors Are Actually Using
Adoption has moved from experiment to standard practice among large investors, and the direction is clear even where the specific tools differ.
JLL research found that by mid-2025, 92% of companies had initiated AI pilots, with real estate data workflows the leading use case. The pattern across large firms tends to follow a few lines:
- Document and lease intelligence: tools that abstract leases and reconcile financials at portfolio scale, the most mature and widely adopted category.
- Market and underwriting analytics: platforms that run comparable analysis and stress-test assumptions against large datasets.
- In-house models on purchased data: the largest firms increasingly build proprietary models on top of licensed data, tailoring the analysis to their own strategy.
The common thread is that these tools sit inside a human-led process rather than replacing it. As industry research stresses, AI in diligence works by compressing timelines and surfacing risk earlier, while experienced investors still make the final call.
The honest caveat is that adoption is uneven. Piloting a tool is not the same as embedding it, and many firms are still working out which workflows genuinely benefit. The gap between a pilot and a dependable part of the process is where much of the current effort sits.
Where Due Diligence Ends and VerbaFlo Begins
Due diligence answers one question, telling you whether an asset is worth acquiring and at what price. It is the work that gets a deal to close. What determines whether the acquired asset delivers the return the model promised is a different discipline that begins once the keys change hands.
A building's projected return depends on how well it performs when occupied. Occupancy, rent collection and resident retention turn an underwriting assumption into a realised return, and those outcomes rest on communication.
That operational layer is where VerbaFlo works. It is a conversational AI platform for residential real estate, handling enquiries, qualification and bookings across voice, chat, WhatsApp and email.
Diligence decides which asset to buy; running the property well protects the return once it is bought. See how it works on your portfolio. Book a demo.
