An AI-fluent brokerage is not the firm with the most proptech subscriptions. It is the firm where a normal Tuesday runs faster because people reach for ChatGPT, Claude, or Gemini on their own real work, get a usable first draft, and check it against the source before anyone trusts it. Fluency lives in habits, not tools. A shop with three chat accounts and a verification reflex is more fluent than a competitor paying for six AI platforms nobody has learned to question. This is the anatomy of the fluent version: what you would see if you walked into a 10-person commercial real estate firm on a working Wednesday, so you can hold the picture against your own firm and see how close you already are.
What “AI-fluent” actually means
Most of the industry is stuck at the pilot stage. JLL’s 2025 Global Real Estate Technology Survey found roughly nine in ten CRE investors and occupiers running AI pilots, yet only about 5% reported hitting all of their program goals. The distance between those two numbers is the distance between owning tools and being fluent. A pilot buys access. Fluency is what a team can do with that access when nobody is watching.
So define it plainly. An AI-fluent brokerage is one where the people who produce written and analytical work — brokers, analysts, property managers, the principal — can complete their real daily tasks faster with an AI tool than without one, and can tell when the output is wrong. Both halves matter. Speed without judgment is a firm one bad hallucinated cap rate away from an embarrassed client. Judgment without speed is a firm that stayed skeptical and never got the hours back. Fluency is the two together, practiced until they are ordinary.
That definition is deliberately falsifiable. “AI literacy” as a vibe is unmeasurable; a broker producing a verified first-pass letter of intent in under fifteen minutes is not. The rest of this piece is the observable evidence that a firm has crossed that line.
The tells: what fluency looks like on a normal day
You can diagnose a firm’s fluency without reading its software invoices. The signs are behavioral.
People reach for the tool unprompted. In a fluent firm, nobody is reminded to use AI. A broker facing a blank LOI opens a chat window the way they would open a template, because it is faster and they trust their own ability to check the result. Unprompted use on real work is the single clearest tell.
First drafts arrive in minutes, not hours. A market write-up that used to eat an afternoon becomes a structured draft the analyst refines over coffee. The tool does not finish the work; it removes the blank-page tax and the mechanical assembly, so the human spends their time on judgment instead of formatting.
There is a shared prompt library, and it grows. Fluent firms accumulate a living document of the prompts that work for their LOIs, their lease summaries, their client email in the firm’s voice — plus the failure notes that taught them what to check. When one person discovers a better way to summarize a lease, the whole firm has it by Friday.
Mistakes are caught early and talked about openly. People trade “watch out, it invented a comp here” the way they trade market gossip. A firm that treats AI errors as normal-and-catchable, rather than as proof the technology is untrustworthy, is a firm that has learned to work with it.
None of these tells requires a specific vendor. They are properties of a trained team, which is why the honest route to them is practice, not procurement — the argument laid out in the case for building fluency before you buy more tools.
Who does what in a fluent firm
A small brokerage is really three firms sharing an office — brokerage, property management, and acquisitions — and fluency shows up differently in each. But the roles that make fluency stick are consistent.
The principal uses it too. In firms that stall, ownership endorses AI and delegates it. In fluent firms, the principal drafts their own investor updates and deal memos with the tool and talks about it in the same breath as everyone else. That participation is what tells the team the skill is how the firm works, not a junior-staff experiment. A principal who models the habit is worth more than any policy.
Someone owns adoption. Fluent firms have a person — often not the most senior, usually the most curious — who keeps the prompt library current, answers the “how do I get it to do X” questions, and notices when a workflow is ripe for a better approach. This is not a full-time job at a 10-person firm; it is a hat somebody wears. Naming that person is one of the highest-return decisions a small firm makes, and it is worth being deliberate about who wears the hat, as covered in the framework for choosing who owns AI adoption.
The skeptic has been converted, not overruled. Every firm has a senior producer who controls real revenue and distrusts the hype. In a fluent firm, that person is not tolerating AI under protest; they were won over, usually by watching a peer produce real work on a live deal faster than they could by hand. Their conversion is a load-bearing part of the culture, because their standards are what keep the firm’s verification honest.
The verification reflex that defines it
If there is one habit that separates a fluent brokerage from a reckless one, it is this: fluent firms treat every AI output as a confident first draft from a fast, occasionally wrong assistant, and they check the parts that decide money before acting on them.
This reflex is not paranoia; it is the entire basis of trust. AI tools produce fluent, plausible text, which means their errors are plausible too — an invented comp, a misread rent-commencement date, a clause reference that points to the wrong section. A fluent broker knows exactly where those failures hide and checks them by habit: every number, date, and party name in an LOI against the term sheet; every extracted clause in a lease summary against the document; every comp in a market write-up against a real source. The check takes a minute and it is non-negotiable.
Fluency, in other words, is disciplined checking as much as it is clever prompting. A firm can write beautiful prompts and still be dangerous if it ships unverified output; a firm with plain prompts and an ironclad verification habit is safe and fast. When you assess your own team, watch for the check, not the keyboard flourish. What tends to fade first after initial training is not prompting skill but this discipline, which is why the durability of the habit is the real test of whether training worked, as traced in what actually sticks in the first thirty days after a workshop.
How a fluent firm handles confidential deal data
Security anxiety kills more small-firm adoption than skepticism does, and it is legitimate: your team handles rent rolls, tenant financials, and deal terms under NDA. A firm that has become fluent has not ignored that risk; it has settled it, which is itself a marker of fluency.
The pattern is consistent. Fluent firms run on business-tier accounts of the major AI providers, where the providers state that inputs are not used to train models by default — a materially different posture from a free consumer login. They set a short, real rule: classify before you paste, anonymize NDA material, and route anything genuinely sensitive through the tools the firm has vetted. And they name one person responsible for the data question so it is not left to each individual’s judgment on a busy afternoon. Verify the specific plan’s data-handling terms at purchase time, because provider terms change; a fluent firm treats that verification as routine, not as a reason to stay on the sidelines.
The tell here is calm. A fluent firm is not frozen by the data question and not cavalier about it. It has three rules it can state in a sentence, and it moves.
What fluency is not
Naming the false positives matters, because the market sells several of them as fluency.
Fluency is not automation. A firm that has paid to automate lease abstraction has bought a machine; whether it is fluent depends on whether its people can judge the machine’s output. Automation is a real and valuable step, but it comes after fluency, not instead of it — a firm that automates a workflow its people cannot quality-check has built something nobody can trust. The full sequence from fluency to automation is the spine of the 90-day plan for making a small firm fluent.
Fluency is not a subscription count. Owning CoStar’s AI features, a Buildout account, and a chat tool is not fluency; it is spend. The firms drowning in proptech they never learned to use are common, and they are not fluent — they are subscribed.
Fluency is not a one-time certificate. A team that completed an online course and went back to old habits is not fluent. Fluency is a standing capability that shows up in this week’s work, which is why the honest measure is always behavioral: what can your people do today that they could not do a quarter ago, without help.
Why the small firm has the edge
Here is the part that should change how a principal feels about the whole exercise. A brokerage of 4–20 people can become fluent firm-wide in a single quarter. An institution cannot.
A large firm answers the AI-literacy question with a center of excellence, an AI lead, a governance committee, and a reskilling program measured in years — because coordinating thousands of people is genuinely hard. A small firm answers it by getting everyone who matters into a few working sessions on real deals and building the habit together. The shared prompt library that a 10-person firm assembles from its own documents becomes a compounding asset: every deal makes it sharper, and it lives in the firm rather than in one person’s head. That combination — full-firm retrainability in a quarter, plus a house knowledge base that grows with every LOI — is a structural advantage the giants cannot copy, and it is the core of the argument in the small CRE firm AI manifesto. Deloitte’s 2026 Commercial Real Estate Outlook, drawn from a survey of more than 850 executives, urges firms to treat AI literacy as a leadership-level priority. Institutions meet that with org charts; a small firm meets it with practice, and gets there faster.
How to tell where your firm stands
Hold your own firm against the tells and score it honestly.
| Marker of fluency | Not yet | Getting there | Fluent |
|---|---|---|---|
| Who uses AI on real work | The curious few | Most, when reminded | Everyone, unprompted |
| Speed on a first-pass LOI | Same as before | Faster, inconsistent | Under 15 min, verified |
| Verification habit | Rare or absent | Sometimes checked | Every number, by reflex |
| Prompt library | None | One person’s notes | Shared and growing |
| The principal | Endorses, delegates | Occasional user | Uses it and models it |
| The senior skeptic | Resisting | Trying it | Converted, sets the standard |
| Deal-data rules | Unspoken | Debated | Three rules, one owner |
A firm that lands in the right-hand column across most rows is fluent, whatever its software bill says. A firm scattered across the first two columns does not need more tools; it needs reps. The honest next question is not which AI to buy — it is whether your firm’s fastest path is a hands-on training session, a specific automation project, or a month of fundamentals first.
Frequently asked questions
What does it mean for a brokerage to be “AI-fluent”?
It means the people who produce written and analytical work can complete their real daily tasks faster with an AI tool than without one, and can tell when the output is wrong. Both halves are required: speed and judgment. A fluent firm shows this in behavior — brokers reach for ChatGPT, Claude, or Gemini unprompted on real deals, produce usable first drafts in minutes, and verify every number against the source before acting.
Is an AI-fluent brokerage just one that buys a lot of AI tools?
No, and this is the most common confusion. Fluency is a property of people and habits, not of software subscriptions. A firm with three business-tier chat accounts and a strong verification habit is more fluent than a competitor paying for six proptech platforms nobody has learned to question. Tool spend and fluency are independent; many heavily-subscribed firms are not fluent at all.
What is the single clearest sign of fluency?
Unprompted use of AI on real work, paired with a verification reflex. When brokers open a chat window to draft an LOI the way they would open a template — without being reminded — and then check every figure against the term sheet by habit, the firm is fluent. If people only use AI when told to, or ship output without checking it, the firm is not there yet.
How is fluency different from automation?
Fluency is people being able to use and judge AI on their own work. Automation is a built system that runs a workflow, like lease abstraction, with less human touch. Automation comes after fluency: a firm that automates a workflow its people cannot quality-check has built something it cannot trust. Get the team fluent first, then automate the workflows where the volume justifies it.
How long does it take a small firm to become fluent?
A brokerage of 4–20 people can reach firm-wide fluency in about a quarter, because you can get everyone who matters into a few working sessions on real deals and build the habit together. This speed is a small-firm advantage; large institutions take years because coordinating thousands of people is far harder. The limiting factor is deliberate practice, not headcount or budget.
Do we need a full-time person to run AI at our firm?
No. At a 10-person firm, owning AI adoption is a hat someone wears, not a full-time role. That person keeps the shared prompt library current, answers “how do I get it to do X” questions, and spots workflows ready for a better approach. Naming that owner is one of the highest-return decisions a small firm makes, but it does not require a new hire.
Is it safe to use AI on confidential deal data?
It can be, with a few rules. Fluent firms use business-tier provider accounts, where inputs are not used to train models by default, rather than free consumer logins; they anonymize NDA material before pasting; and they name one person responsible for the data question. Verify the specific plan’s data-handling terms at purchase, because provider terms change. The concern is real but solvable, and solving it is itself a marker of fluency.
What does becoming fluent cost?
The most common on-ramp is a hands-on team workshop, which runs roughly $2,000–$15,000 in the current market depending on depth, length, team size, and format. That builds the habit; it does not require buying new software beyond business-tier chat accounts. Custom automation, if a firm later wants it, is a separate project at a different scale, so the right number depends on your firm’s starting point.
Where to start
The picture above is a target, not a verdict. Most small firms land somewhere in the middle — fluent in one corner, absent in another — and the useful question is which single move closes the most distance. That read is exactly what a free AI-readiness assessment produces: a short working session that maps your firm’s real workflows, finds where the hours actually go, and returns a plan for whether a hands-on training session, a specific automation project, or a month on fundamentals is your fastest path to a fluent Tuesday. Book a free AI-readiness assessment if you want that read before you spend a dollar.
Arthur Wandzel
