When a small commercial real estate firm decides to “do something with AI,” the first move is almost always the wrong one: someone opens a browser and starts shopping for software. A lease-abstraction product, an AI-enabled CRM, a deal-screening platform — pick a task, buy a tool, book a demo. It feels like progress because it looks like a decision. The case for training before tooling is the argument that this order is backwards, and that the firms getting real value are the ones that make their people fluent with a general assistant first and buy specialized software second. This is not an anti-tool position. It is a sequencing position, and the sequence matters because the two purchases are not independent: how fluent your team is decides whether the tool you buy next earns its keep or joins the pile of subscriptions nobody opens. This piece makes that case, names the narrow exceptions where a tool really does come first, and lays out what to spend on in what order.
What “training before tooling” actually means
Training before tooling means making your team fluent with a general AI assistant on your real work before you commit budget to any purpose-built proptech product. Fluency here is a concrete, practiced skill: a broker who can draft a letter of intent, summarize a lease, or write a market update with ChatGPT, Claude, or Microsoft Copilot and trust the result after a quick check. It is not a certificate or an awareness session. It is the ability to get useful work out of a general tool your firm can stand up in an afternoon.
The distinction that trips firms up is between a general assistant and a specialized tool. A general assistant is the horizontal layer — it drafts, summarizes, reformats, and answers across every document type your firm touches. A specialized tool is vertical — a lease-abstraction engine, an underwriting copilot, a CRM with AI features bolted on — built to do one job with more structure and less prompting. Both have a place. The argument is only about order: fluency on the general layer is the foundation that makes every vertical tool you buy afterward more likely to stick.
Read that way, “training before tooling” is not a rejection of software. It is a claim about return on the software you will eventually buy. The same $30,000 lease tool produces a very different outcome in a firm whose team already thinks in terms of what AI can and cannot do than in a firm where it lands cold. The sequence is the strategy.
Why small firms reach for the tool first
The tool-first reflex is not stupidity; it is the shape of how firms are used to buying. Software is a line item you can approve, expense, and point to. Training feels softer, harder to measure, and easier to defer. Faced with a choice between a purchase order and a behavior change, most principals sign the purchase order — it closes the loop faster.
Three forces push the same way. Vendor marketing makes the tool look like the whole answer; every proptech demo ends with “and that’s why you need this,” never “first your team should get fluent on something cheaper.” The procurement instinct treats a subscription as a decision made and training as an ongoing obligation, so the subscription wins. And the demo effect is real: a slick walkthrough shows the finished result without the fluency it quietly assumes the operator already has.
The result is a familiar pattern. A firm buys a specialized tool, a champion uses it for two weeks, adoption thins, the renewal comes up, and nobody can say whether it paid for itself. The tool was not bad. It arrived before the firm could use it. Our look at what actually sticks after a dozen AI workshops traces the same failure from the training side: capability that is not reinforced does not survive contact with a busy week.
Five reasons fluency has to come first
The case rests on five arguments. Each stands alone; together they are hard to answer with “let’s just buy the tool.”
You can’t evaluate a tool you can’t use
You cannot judge a specialized AI product before your team can use a general one. Evaluating a lease-abstraction tool means knowing what “good” output looks like, where these systems break, what a reasonable error rate is, and how much review the output still needs. A team that has spent a month summarizing leases with a general assistant knows all of that from experience and can run a demo like a skeptic. A team that has not is evaluating on vibes and vendor claims — the worst possible position for a firm that has been burned by proptech subscriptions before. Fluency is the prerequisite for good procurement, not a luxury you add afterward.
Seats bought before fluency become shelfware
The default fate of a tool bought before fluency is unused seats. This is not a small-firm quirk; it is the dominant outcome across the industry. JLL’s 2025 Global Real Estate Technology Survey found that roughly nine in ten CRE firms are piloting or adopting AI, yet only about 5% report hitting all their program goals. The gap between those two numbers is mostly people who were handed access to something they were never made fluent on. Buying the tool is the easy 10%; using it is the missing 90%. Fluency first inverts the odds — a team that already reaches for AI daily will actually open the specialized tool you buy them.
A general assistant already covers most of day one
For most of what a small CRE firm wants on day one, a general assistant is already enough. Drafting LOIs and cover emails, summarizing a lease or an offering memorandum, turning messy notes into a market write-up, cleaning up a rent-roll narrative — a business-tier general assistant handles all of it with good prompting and no procurement cycle. You do not need a specialized product to capture the first wave of value; you need a fluent team and one subscription. Proving that value on cheap, flexible tooling is how you learn what specialized problem is actually worth paying to solve.
Fluency is portable; tool skills are not
Fluency transfers; tool-specific skill does not. The reasoning habits your team builds on a general assistant — how to frame a task, how to check output, when to trust it and when not to — apply to every future tool and every new task. The skills required to drive one vendor’s particular interface evaporate the moment you switch vendors or the product redesigns its workflow. Investing in portable fluency first means every tool you adopt later plugs into a team that already knows how to think with AI, rather than starting the learning curve over with each new subscription.
Training changes what you end up buying
Fluency changes the shopping list, usually for the better. Teams that get fluent routinely discover that two of the four tools they were about to buy are unnecessary — a general assistant already covers the job — and that the one tool they do need is different from the one the demo sold them. They also spot the genuine gaps more clearly: the sustained, high-volume task where prompting by hand stops scaling and a purpose-built system earns its price. Training first does not just improve how you use tools; it sharpens which tools you buy at all, which is where the real money is saved.
The evidence: pilots stall on people, not software
The broader data says the bottleneck in AI adoption is human, not technical. The World Economic Forum’s Future of Jobs Report 2025 estimates that 39% of workers’ core skills will change by 2030, with AI and data skills the fastest-growing category — a signal that the durable investment is in people who can keep up, not in a single product snapshot. Deloitte’s 2026 Commercial Real Estate Outlook names the downstream symptom directly: “AI pilot fatigue,” the state firms reach when they launched initiatives — usually tool purchases — without a plan to sustain the capability behind them.
Put the numbers together and the pattern is unambiguous. Firms are buying and piloting at scale; very few are getting the full return; and the reports keep pointing at skills and sustainment rather than software as the missing piece. Tooling is not where these initiatives fail. People-readiness is.
The upside of fluency, meanwhile, is well documented and specific to the work CRE teams do. A Science study found professionals completed mid-level writing tasks about 40% faster with AI assistance, and a field study of more than 5,000 workers found average productivity gains of 14%, rising to 34% for less-experienced staff. CRE drafting and document work sits squarely in that band — and those gains come from fluency applied to daily work, not from a specialized product.
What comes first, concretely
Sequencing training ahead of tooling is not vague advice; it is a concrete order of operations a small firm can run in a quarter.
- Stand up one business-tier assistant. Give the whole team ChatGPT Business, Claude Team, or Microsoft Copilot — roughly $20 to $30 per user per month. The business tier matters because it keeps confidential deal data out of model training under the vendor’s terms, which the free consumer tiers do not guarantee. One assistant firm-wide is the entire day-one stack.
- Run one fluency session anchored to your own deals. Not a generic intro-to-AI webinar — a hands-on session using your executed LOIs, real lease summaries, and actual market write-ups, so the skill survives past Friday.
- Fund reinforcement, not hope. A maintained prompt library, a short monthly demo, and a named internal champion who owns the habit. This is the step firms skip and the reason sessions decay.
- Only then scope tooling. With a fluent team, evaluate specialized products against a problem you now understand from the inside — and buy the one or two that clear the bar, not the four the demos sold you.
For the full 90-day version of this sequence — who does what, in what order — our CRE AI training playbook lays it out step by step, and the AI adoption framework for small CRE firms covers how to structure the rollout so it holds. The point of the order is not to delay tooling forever. It is to arrive at the tooling decision with a team that can make it well.
When tooling genuinely comes first
Honesty about the exceptions is what separates a sequencing principle from dogma. There are real cases where a purpose-built tool should lead.
The clearest is sustained, high-volume document work. A property-management firm abstracting hundreds of leases a month, or an acquisitions shop processing large offering-memorandum stacks on a deadline, hits a ceiling where prompting a general assistant by hand stops scaling. There, a purpose-built lease-abstraction or document-intelligence system with structured extraction and review workflows earns its price on volume alone — and waiting on fluency would leave measurable throughput on the table.
Two more cases qualify. Compliance-driven pipelines — where output must be auditable, versioned, and consistent across a team — often need the guardrails a specialized product provides rather than the free-form output of a general chat tool. And deep system-of-record integration — AI features inside the CRM or accounting platform your firm already runs on — can be worth switching on early because the data is already there.
Even here the sequencing logic holds in miniature: the people operating the specialized tool still need to understand what good AI output looks like to run it well. The exception narrows the argument; it does not overturn it. If your firm is not doing sustained high-volume work, has no compliance mandate forcing the issue, and is reaching for a tool mainly because shopping feels like progress, fluency comes first.
Sizing the two spends
Framed as two purchases, the order makes budgeting sense as well as adoption sense. A facilitated AI fluency session for a small team runs roughly $2,000 to $15,000 depending on format, with business-tier assistant access adding about $20 to $30 per user per month. That is the training layer — modest, fast to stand up, and useful across every task in the firm.
Custom automation and specialized proptech are the larger, later commitment — a scoped custom automation project runs roughly $25,000 to $150,000 depending on complexity, with specialized SaaS licensing on top. Spending the smaller training amount first is what protects the larger tooling amount from becoming shelfware: cheap insurance on an expensive purchase.
There is a competitive argument underneath the sequence, too. A lean firm that gets fluent can capture the first wave of AI value with one subscription and a trained team, then buy specialized tools deliberately rather than reflexively — the same out-execute-the-giants logic laid out in the small-firm AI manifesto. For how to fund the training layer as a continuous line rather than a one-time course, see our piece on rethinking the training budget in the AI era. The firms that win here are not the ones that bought the most software. They are the ones that could use what they bought.
FAQ
What does “training before tooling” mean for a CRE firm?
It means making your team fluent with a general AI assistant on your real work before you buy specialized proptech software. Fluency is the practiced ability to draft LOIs, summarize leases, and write market updates with ChatGPT, Claude, or Microsoft Copilot and trust the result after a quick check. Training before tooling is a sequencing argument, not an anti-tool one: get the people ready first, then buy the vertical tools that clear a bar your fluent team can now judge.
Why should a small firm train its team before buying AI tools?
Because how fluent your team is decides whether the tool you buy next gets used or ignored. JLL found roughly nine in ten CRE firms piloting AI but only about 5% hitting all their program goals — the gap is mostly people handed access they were never made fluent on. A fluent team can evaluate specialized tools like a skeptic, actually adopt what you buy, and often discover they need fewer tools than the demos suggested. Training first raises the return on every tool bought later.
Isn’t buying a purpose-built tool faster than training people?
It is faster to purchase and slower to pay off. A specialized tool bought before fluency typically sees a champion use it for a couple of weeks, then adoption thins and the renewal arrives with no clear answer on whether it paid for itself. Standing up one business-tier assistant and running a deal-anchored fluency session takes weeks, not months, and makes every later tool purchase land better.
Can a general assistant replace specialized proptech tools?
For most day-one CRE work, yes — drafting, summarizing, reformatting, and market write-ups run well on a general assistant with good prompting and no procurement cycle. Specialized tools earn their place on sustained, high-volume, or compliance-driven work where prompting by hand stops scaling. The right approach is to capture the first wave of value on a general assistant, learn from it which specialized problem is actually worth paying to solve, and buy the vertical tool for that specific job.
When should a CRE firm buy the specialized tool first?
When it does sustained, high-volume document work — hundreds of lease abstractions a month or large offering-memorandum stacks on deadline — where a purpose-built system beats hand-prompting on throughput. Compliance-driven pipelines that need auditable, consistent output, and AI features inside a system of record you already run on, also justify buying early. Outside those cases, if you are reaching for a tool mainly because shopping feels like progress, fluency should come first.
How much does training before tooling cost versus the tools themselves?
Training is the smaller spend: a facilitated fluency session runs roughly $2,000 to $15,000 depending on format, plus about $20 to $30 per user per month for business-tier assistant access. Specialized software and custom automation are the larger, later commitment — a scoped custom automation project runs roughly $25,000 to $150,000, plus SaaS licensing. Spending the smaller training amount first is what keeps the larger tooling amount from turning into unused seats.
Does training before tooling apply to property management and acquisitions too?
Yes. Property managers get fluent on lease abstraction and tenant correspondence, acquisitions teams on summarizing offering memoranda and drafting investor updates, and brokers on LOIs and market write-ups. The document types differ but the compressible core — language and lookup work — is the same across the firm, and it runs on a general assistant before any specialized tool.
What happens if we skip training and just buy the tools?
You usually end up paying for software your team does not use and cannot evaluate. Without fluency, the firm can neither judge whether a tool’s output is good nor adopt it past an initial champion, which is the “AI pilot fatigue” Deloitte’s 2026 outlook describes — initiatives launched without a plan to sustain the capability behind them. The tool is rarely the problem; it arrived before the firm could use it. Training first is the cheapest way to avoid that outcome.
How do we justify training spend to a partner who wants a tool?
Frame it as protecting the tool purchase, not competing with it. The training layer is a small, fast spend that raises the utilization — and the return — of every larger tooling investment that follows, and it lets the firm buy those tools with judgment instead of on vendor claims. A fluent team also tends to buy fewer, better-chosen tools, so training first often reduces total software spend rather than adding to it.
Key takeaways
- Training before tooling is a sequencing argument: make your team fluent with a general assistant first, then buy specialized software — the two purchases are not independent, and fluency decides whether the tool pays off.
- You cannot evaluate a specialized tool your team cannot use; fluency is the prerequisite for good procurement, not an add-on after it.
- Seats bought before fluency default to shelfware — JLL found ~9 in 10 CRE firms piloting AI but only ~5% hitting all their goals; fluency first inverts those odds.
- A business-tier general assistant covers most day-one CRE work — LOIs, lease summaries, market write-ups — with no procurement cycle, and fluency built on it transfers to every tool you buy later.
- The exceptions are narrow and real: sustained high-volume document work, compliance-driven pipelines, and deep system-of-record integration can justify buying a tool first — but outside those, fluency leads.
Want to know which order is right for your firm — and which tools you actually need? A short conversation about your team, your workflows, and where you are starting from will answer it far better than any demo. Book your free AI-readiness assessment → and we will map what fluency would cost, what to buy, and in what order.
Arthur Wandzel
