State of AI in CRE 2026: What Firms Are Actually Doing (and Where Most Still Aren't) | NextAutomation

This is our annual read on the state of AI in commercial real estate for 2026, and the short version is uncomfortable in both directions. Nearly every firm is now piloting AI, so being "in" on AI is no longer a differentiator. But almost nobody has finished: a tiny minority have gotten AI to actually do the job they bought it for. If you feel behind, you are behind the pilots, not behind the results. Almost no one is at the finish line yet.

Most "state of AI" reports are built from a survey and a vendor's marketing goals. This one is built differently. It is led by what CRE operators actually asked us for, then reads the independent industry surveys as context on top. Over the last several months, 8,092 distinct operator sessions viewed or requested our free CRE-AI templates, six real acquisitions and investment principals told us on recorded calls exactly what they are already running in ChatGPT and Claude, and 150 messages came in through our site chat asking the same handful of questions. That first-party demand is the spine of this report. The published industry surveys from JLL, Deloitte, Altus Group, and EliseAI are the field context, attributed and dated where we cite them.

One honest note before the data: this is not a survey and we are not going to dress it up as one. It is real demand and real conversations, with the sampling bias stated plainly. We publish the honest sample size because in a category this full of inflated numbers, the honest number is the more useful one. The methodology box near the end tells you exactly what this data is and is not.

Where are operators actually pointing AI? Sourcing first. That is the single clearest signal in our first-party data, and it runs against the assumption that underwriting is the obvious first automation. Here is the full read.

The Executive Read: Everyone Is Piloting, Almost Nobody Has Finished

If you take one thing from this report, take this: the gap that matters in 2026 is not pilot versus no-pilot. It is pilot versus finished. Per JLL's 2025 Global Real Estate Technology Survey, published October 28, 2025 and drawn from more than 1,500 decision-makers across 16 markets, 88% of investors and owners have started piloting AI (92% of occupiers), up from under 5% in July 2023. In under three years, piloting AI went from a rounding error to near-universal. That is why simply "doing AI" no longer sets you apart.

The finish line is where almost everyone still stands. In that same JLL survey, only 5% of firms report achieving all of their AI goals; 47% met just two or three of them. Firms are running roughly five pilots at once, JLL found, across 56 identified use cases, and 87% say their technology budgets rose because of AI. The spend is real, the pilots are real, and the completed outcomes are rare. JLL also reports that more than 60% of investors consider themselves unprepared, strategically, organizationally, and technically, to scale AI beyond pilots. So the honest state of the industry is: wide adoption, shallow completion, and a readiness wall between the two.

Our own first-party demand data tells you where operators are trying to cross that wall, and the surprise is the direction. The loudest, fastest-growing demand is not for underwriting automation. It is for sourcing: finding the deal and finding the owner before it hits the market. Below are the five cuts from our own data, then the field context, then the honest methodology, then a way to self-locate.

Cut 1: The Demand Census, Read Honestly

Between February 27 and July 9, 2026, 8,092 distinct operator sessions viewed or requested our library of 122 free CRE-AI templates, generating 10,293 views. Those are the operators raising their hand for AI they can actually use: prompts, agents, and templates for real deal work. We say viewed or requested rather than downloaded on purpose. The count is genuine hand-raising, and we would rather describe it precisely than inflate it.

Two caveats ship with this number, because leaving them out would make it a worse signal, not a better one. First, this is a LinkedIn-audience-weighted census: 57% of the views came through LinkedIn distribution, so this reflects the operators in that audience, not the whole industry. Second, and more important, this is not a month-over-month growth curve. March and April together carried roughly 5,800 of the 8,092 lifetime sessions. That is a distribution spike followed by a long tail, the classic shape of a piece of content that traveled, not evidence that adoption is accelerating on its own. Anyone who hands you a clean up-and-to-the-right AI adoption curve from data like this is selling the curve. Demand concentrates in a spike, then trails.

Read correctly, the census tells you one durable thing: when you put genuinely useful CRE-AI templates in front of operators, thousands reach for them. The appetite is not the question anymore. What they reach for is the question, and that is Cut 2.

Cut 2: Investors and Acquisitions Outweigh Brokers Roughly 4.5 to 1

Sorted by who is asking, the demand skews hard toward the buy side. Investor and acquisitions sessions outnumbered broker sessions roughly 4.5 to 1: 4,853 versus 1,066 distinct sessions. Developers registered 311 sessions, and the institutional and LP bucket showed 48, real but tiny, which we read as under-distribution to that audience rather than proven under-demand.

A note on how to read those numbers: we publish the ratio, not percentage shares, because these buckets are topic-affinity groupings, not a clean partition. A single operator can show up in more than one bucket, so shares would overstate precision we do not have. The ratio is the honest, load-bearing figure: buy-side operators are reaching for practical CRE-AI at several times the rate of the brokerage side.

The more interesting finding sits inside that buy-side demand. When acquisitions teams tell us what they want AI for, they name sourcing before underwriting. In our recorded calls, off-market sourcing and owner-finding came up in 13 distinct real conversations; underwriting and pro-forma automation came up in 6. Underwriting matters, but it is not the first thing operators reach for. The first thing is finding the deal. We unpack that in how to find off-market properties, and Cut 3 shows just how fast that demand is moving.

Cut 3: The Off-Market Surge

The fastest-moving demand in our entire library is off-market sourcing, and it is not close. The off-market operating system is the hottest-velocity asset we have ever shipped, running at 18.4 sessions per day across 184 sessions, 100% of them in the last 30 days. And the off-market persona bucket went from 0 to 182 sessions in July alone. This is not a slow build. It is a step change concentrated in the most recent weeks of the census.

Put Cut 2 and Cut 3 together and the operator priority for 2026 is legible: buy-side firms want AI to help them find deals and owners before the deal is competitive. We unpack that buy-side signal on its own in what CRE acquisitions teams actually want from AI sourcing. That is a very different first move than the industry narrative of "AI will speed up your underwriting." It is happening, but it is second.

Cut 4: What Operators Are Actually Stuck On

Demand tells you what operators want. The stuck-points tell you why they have not gotten there, and we give them a full treatment in what CRE firms are actually stuck on with AI. We measure this from two separate corpora, and we keep them separate on purpose: they are different sources with different depth, and summing them would invent a total that does not exist. Read each on its own terms.

From our recorded calls (distinct real conversations, reported as floors): off-market sourcing led at 13 calls, followed by underwriting and pro-forma automation at 6, and operators already self-serving in ChatGPT or Claude and hitting a ceiling at 6. Then reliability and hallucination concerns at 3, workflow decomposition at 4, data security at 3, AI stigma or "I am not a tech person" at 3, build-versus-buy reasoning at 3, and team adoption or firm-wide rollout at 2. These are floors: distinct conversations where the theme was explicit, not a guess at how many people quietly worry about each thing.

From our site chat (150 user messages across 70 sessions): the ranking shifts, because the questions people type into a chat box are more practical and less strategic. The largest theme by a wide margin was adoption and how-it-works, or "how does this actually land in my business," at 38 messages. Pricing came next at 19, off-market and deal-sourcing intent at 13, build-versus-buy and customization at 11, liability and reliability at 6, real-AI-versus-just-automation at 5, and data security at 3.

Two honest reads across both corpora. First, sourcing is not just the top demand, it is also a top stuck-point: operators want it and have not solved it. Second, the two biggest practical blockers in the chat corpus are not skepticism about AI. They are "how does this actually land in my business" and "what does it cost." The reliability and "is this real AI or just automation" objections are real but smaller than the industry assumes. If you have been holding off because you thought the room was full of AI skeptics, the data says otherwise. The room is full of operators who believe it works and cannot see how to land it in their own shop. Two of those questions get their own treatment in build vs buy AI for commercial real estate and real AI or just automation in CRE.

Cut 5: The Adoption Frontier, in Operators' Own Words

Here is the cut that should change how you think about being behind. Six distinct real ICP operators told us, on recorded calls, that they are already self-serving with ChatGPT and Claude before ever buying anything. They have written their own prompts, built their own templates, bought Claude Max plans, and wired up skills and projects. And they all hit the same ceiling: reliability and completeness. Not "does AI work," but "can I trust it end to end without checking every line."

Three de-identified verbatims, lightly trimmed for length and stripped of any identifying detail:

• "We created a prompt where I can now just put an OM into Claude, and it'll spit out an LOI for me. I'm older, right? I'm early 40s, I'm new to this AI thing, and it's pretty cool." A principal at a fully-integrated industrial investment firm.
• "I've pretty much implemented everything and I've been able to leverage AI to my advantage. I have templates for different deal types, my virtual assistant uploads the OM, and gets it to a point where the data is filled in." A solo investor who self-built his stack and is now stuck at reliability.
• "I'm using them a good amount, but I guess my wish would be for something that is more of the complete process as much as possible." A CRE investor evaluating underwriting AI.

Read those again and notice who is talking. Not laggards. These are operators running OM-to-LOI prompts in raw Claude today. The person you are worried about outrunning you is not sitting on the sidelines waiting for a vendor to save them. They are already piloting, in their own account, and the thing holding them back is the same thing holding you back: getting AI to be reliable and complete enough to trust without babysitting it. That is the real frontier in 2026. Not access to AI, which is universal, but crossing from "it mostly works in my chat window" to "it runs the process I can stake a decision on."

The Field Context: What the Industry Surveys Say

Our first-party data tells you where operators are pointing AI and what stops them. The published industry surveys tell you the shape of the whole field. We cite them as context, attributed and dated, never as our own numbers.

The headline from JLL's 2025 Global Real Estate Technology Survey (published October 28, 2025, more than 1,500 decision-makers across 16 markets) matches what we see: adoption is wide and shallow. 88% of investors and owners are piloting AI, up from under 5% in July 2023, yet only 5% report hitting all their AI goals and 47% met just two or three. Firms run roughly five pilots at once across 56 use cases, 87% report technology budgets up because of AI, and more than 60% consider themselves unprepared to scale beyond pilots. Appetite is not the constraint; readiness is.

Deloitte's 2026 Commercial Real Estate Outlook (published September 29, 2025, more than 850 C-level respondents across 13 countries) points at the same wall from the demand side: roughly 75% of respondents plan to increase real estate investment over the next 12 to 18 months, and Deloitte frames reliable data and application readiness, not enthusiasm, as the path to getting value from AI. The money is coming; the readiness has to catch up to it.

Altus Group's US sentiment survey (Q4 2023, published December 2023, 197 respondents across 51 firms) adds the two divides that explain a lot of the "am I behind" anxiety. 72% of respondents said AI will benefit CRE professionals, but the practical-application view splits sharply by firm size: roughly 90% of firms with more than $5 billion in CRE exposure saw practical AI applications, versus roughly 50% at firms under $1 billion. And it splits by tenure: AI positivity ran near 80% for professionals with under 15 years in the industry versus roughly 60% for those with more than 20. If you are a smaller or longer-tenured operator feeling out of step, that feeling is documented, not imagined. It is also not destiny, as our own frontier cut shows early-40s principals already running Claude in production.

One vertical has its own harder numbers, and it is worth naming separately because the source is a vendor. In multifamily specifically, EliseAI's State of AI in Multifamily report (published September 26, 2025, 280 executives) found that 92% of multifamily operators have implemented AI, 32% think their competitors are moving faster, and 72% fear that slow adoption will hurt NOI. That is a multifamily-only, vendor-run survey, so we keep it in the multifamily lane and name EliseAI as the source rather than generalizing it across asset classes. Inside multifamily, though, it sharpens the same picture: adoption is near-universal and the anxiety about pace is real.

How This Report Is Made (and Why We Publish Honest Numbers)

So, Are You Behind?

Locate yourself honestly against the data, not against the noise. Our CRE AI self-check walks you through the same read on your own firm.

• If you have not started piloting at all: you are genuinely behind on the starting line. 88% of investors and owners are already piloting (JLL, October 2025). But starting is cheap now, and the operators ahead of you are stuck at the same reliability ceiling you would hit, so the gap is smaller than it looks.
• If you are piloting but nothing has stuck: you are exactly average, and that is not an insult. Only 5% of firms have hit all their AI goals (JLL, October 2025). The wall you are at, going from "it works in a chat window" to "it runs a process I can trust," is where roughly everyone is stuck. Being here is not falling behind. Staying here is.
• If you are already self-serving in ChatGPT or Claude: you are ahead of the field on adoption and stuck at the same frontier as the six operators in Cut 5, reliability and completeness. The next move is not more prompting. It is turning the prompt into a process you and your team can rely on without checking every line.

The honest conclusion of the 2026 state of AI in commercial real estate is that the race is not "adopt AI," which nearly everyone has, but "finish it," which almost no one has. The firms that pull ahead this year are the ones that cross from pilot to reliable, repeatable process, and do it on the workflow that matters most to their buy side: sourcing first, then underwriting.