The License Is Paid. Nobody Knows How to Use It.

Every year, ICSC Las Vegas is where commercial real estate takes its own temperature. Deals get done, relationships get renewed, and the industry's collective anxiety finds a topic to attach itself to. This year, the topic was inescapable: artificial intelligence — specifically the growing gap between what firms are paying for and what they are actually getting out of it.

Across booth conversations, panel corridors, and after-hours dinners, a remarkably consistent story kept surfacing. Brokerage firms, institutional owners, retail asset managers, and legal teams had all made the move. They had purchased a Claude Team or Enterprise license, or signed up for Microsoft 365 Copilot. The budget decision had been made. The ROI case had been approved.

And then — nothing. Or near nothing.

Not because the tools are bad. But because no one had been given the framework, the workflows, or the protected time to build real skills around them. The license sat open in a browser tab, occasionally used to polish an email or summarize a document, while the work that could transform a practice — deal analysis, lease abstraction, risk flagging, market briefs — kept getting done the old way.

The Four Personas. One Shared Problem.

The Broker

Running at full speed — pitches, site tours, deal negotiation, client reporting. There is no slack in the calendar for experimentation. AI sounds useful but trying to figure it out is one more thing that doesn't make next month's quota.

"I know I should be using this. I just can't stop to learn how."

The Asset Manager

Managing a portfolio with complexity that compounds quarterly. Lease variance, tenant health, NOI tracking, refinance decisions — the analysis workload is relentless. They see AI as a potential analyst but can't operationalize the handoff.

"I need it to do work, not just answer questions."

The Property Manager

Managing a portfolio with complexity that compounds quarterly. Lease variance, tenant health, NOI tracking, CAM audit decisions — the analysis workload is relentless. They see AI as a potential analyst but can't operationalize the handoff.

"I need it to do work, not just answer questions."

The In-House Lawyer

Reviewing lease after lease, flagging risk, managing outside counsel, handling compliance. The document burden is extreme. The concern about AI accuracy — and liability — is real. They need guardrails before they can let go.

"I don't trust it enough to rely on it yet. I don't know where to start."

These are not technology skeptics. They are professionals with full plates and genuine curiosity, looking for a path that doesn't require them to stop doing their job in order to learn the new one.

The pain was not uniform in its detail, but it was uniform in its shape. Four distinct groups of CRE professionals all arrived at the same wall from different directions.

One of the most significant findings from ICSC Las Vegas was how widespread and explicit the resistance to new SaaS has become. This is no longer a quiet preference — it is an organizational policy in many firms. The message from procurement, from IT, from practice group leaders is the same: no new tools.

The stack is overloaded. The integrations are fragile. The per-seat costs compound. The onboarding cycles take months. The promised productivity gains often never materialize because adoption stays shallow. CRE firms have been through this cycle enough times to recognize it.

What "Agentic Workflows" Actually Means for CRE

The word "agentic" gets thrown around a lot in AI circles. In the CRE context it means something specific and immediately recognizable: a workflow where the AI doesn't just answer a question, but executes a sequence of tasks — pulling, analyzing, formatting, and delivering output — with minimal human prompting at each step.

For a broker, that looks like a Claude-powered brief that pulls comparable deals, synthesizes market signals, and produces a formatted pitch narrative from a single instruction. For an asset manager, it's a portfolio health report that ingests rent roll data, flags variances against budget, and surfaces the top five items requiring attention. For in-house counsel, it's a lease abstraction that extracts every material clause, maps deviations from standard positions, and outputs a risk-ranked memo ready for review.

None of this requires custom software. None of it requires a new SaaS platform. It requires well-designed prompts, structured workflows, and people who know how to build and run them. That is the skill gap ICSC made visible — and it is entirely closeable.

Why the Gap Persists

The most honest answer is time. Learning to build effective AI workflows is not hard, but it is not zero-effort either. It requires protected hours — time carved out from billable work, client calls, and deal flow — to experiment, iterate, and internalize new patterns. In commercial real estate, that time is almost never available organically.

The second answer is risk tolerance. Professionals in CRE carry significant accountability. A lease missed, a risk flag overlooked, a deal number that is off — the consequences are real. Until someone trustworthy has validated a workflow in a CRE context, there is rational reluctance to rely on it. That validation layer doesn't exist inside most firms today.

The third answer is that nobody has shown them what good looks like in their specific role. Generic AI tutorials and vendor webinars talk about AI in the abstract. They do not demonstrate what a Claude-powered lease risk assessment actually looks like, how an asset manager builds a portfolio brief prompt from scratch, or what safe adoption means when a junior analyst is using AI to support work that a senior partner will sign off on.

What the Industry Is Actually Asking For

Stripped of jargon, here is what CRE professionals at every level said they wanted when the conversation got honest.

Brokers want to walk away with three or four repeatable prompt sequences that cut deal prep time in half — and the confidence to run them without double-checking every output. They want to see a live demo of what a properly prompted AI can produce for a pitch, a comp analysis, and a market brief.

Asset managers want a portfolio monitoring workflow they built themselves and understand completely — so when a number looks wrong they know why, and when a flag fires they trust it. They want AI as an analyst, not a black box.

Property managers want relief from the documentation burden — tenant correspondence, CAM reconciliation summaries, vendor communications — produced fast, in their voice, without adding a new system to their stack.

In-house lawyers want a safe lease workflows with clear guardrails — a process they can defend to their GC, that flags what needs human review and passes only what is genuinely low-risk. They want to speed up the funnel without increasing exposure.

All four groups want the same thing underneath: skills they own, workflows they control, and a session structured well enough that they can build them without disrupting the business to do it.

Half a Day. Real Workflows. No New Software.

The half-day intensive was designed specifically for this moment — for firms that have the license, have the intent, and are ready to stop talking about AI and start building with it.

It is not a webinar. It is not a vendor pitch. It is a working session: role-specific, hands-on, and structured to produce actual deliverables that participants use the following Monday. By the end, every person in the room has built at least one agentic workflow in their own domain, understands the prompting principles behind it, and has a clear roadmap for what to build next.

No new tools. No new integrations. No new budget request. Just a structured half-day that unlocks what the firm already paid for.

The Takeaway from Las Vegas

ICSC Las Vegas confirmed what many in the industry suspected but hadn't yet named out loud: the adoption problem in commercial real estate AI is not a technology problem. The technology is good, the licenses are purchased, and the ROI case is understood. The problem is structured skill-building at the practitioner level — and that is a solvable problem.

The firms that figure this out in the next twelve months will not just save time. They will build a durable capability advantage — a team that thinks in workflows, operates with leverage, and owns its intelligence rather than renting it from a vendor's platform. In a market where execution speed and analytical depth increasingly determine who wins the mandate, that is not a marginal edge.

It is the edge.