Short answer
An "AI utility belt" — a loose kit of chat tools for drafting, summarizing, and prospecting — helps an individual write faster but leaves an operating company with drafts nobody approved and numbers nobody checked. A CRE firm is better served by one governed system that stages actions on your rent roll and leasing tracker and requires a human to approve them.
What is an "AI utility belt" for CRE, and why is it appealing?
The "utility belt" is the idea that you assemble a handful of general AI tools — ChatGPT for emails, a summarizer for leases, a scraper for prospecting — and reach for whichever one fits the task. It is appealing because each tool is cheap, works today, and needs no rollout. For a solo broker writing a follow-up, that is genuinely useful.
The problem starts when the firm, not the individual, is the unit of work.
- Each tool holds its own scrap of context; none knows your rent roll (the current schedule of tenants, rents, and lease terms).
- Output lands in a chat window, not in your system of record (the one place your deals, leases, and contacts actually live).
- Nobody signs off. A tool that drafts is also a tool that can send.
Why does a pile of point tools break down for an operating company?
Because generation is now the cheap part; judgment and accountability are the work. A utility belt multiplies drafts and answers without multiplying the checking. According to a 2024 McKinsey survey, 65% of organizations regularly use generative AI, yet most report they have not put governance around inaccuracy or oversight. That gap is where an operating company gets hurt.
Concretely, six disconnected tools create six problems:
- Six copies of the truth. Each tool answers from whatever you pasted in, so a rent figure in one thread contradicts the lease in another.
- No handoff. A summary in one window can't become a task in your tracker without someone retyping it.
- No audit trail. When a wrong NNN figure — a triple-net lease where the tenant pays taxes, insurance, and maintenance — reaches a tenant, you can't reconstruct who approved it.
- Silent commits. Tools that draft and send in one step turn a rough answer into a sent email.
What should a CRE firm use instead?
One governed desk instead of a belt of tools. Vantrow's guiding principle is propose, never commit: the system stages an action — a broker follow-up, a lease-expiration alert, a rent-roll update — and a person approves it before anything leaves the building. Everything lands on an audit trail. You keep the speed of generation and add the accountability an operating company runs on.
The practical difference:
- Utility belt: a tool drafts an LOI response and can send it.
- Governed desk: the desk drafts the LOI response, attaches the source lease clause, and waits for a human to approve.
An LOI (letter of intent) is the non-binding term sheet that precedes a lease or sale — exactly the kind of document you never want a tool sending on its own.
How does a governed desk actually work day to day?
The desk sits on top of your existing artifacts — the leasing tracker, the rent roll, the contact list — and treats each AI action as a proposal a person accepts or rejects.
- A broker texts an update from a site visit; the desk drafts the tracker entry, a person confirms it.
- A lease is 180 days from expiration; the desk stages a renewal follow-up, an asset manager approves it.
- A prospect list comes in; the desk gathers freely but sends carefully — outbound waits for sign-off.
Every proposal carries its source and its approver. That is the difference between speed you can defend and speed you have to apologize for.
Where is the "utility belt" approach actually fine?
For low-stakes, single-person, throwaway work. Drafting a first pass of a marketing blurb, brainstorming a subject line, or summarizing an article you'll read anyway — reach for a general tool. The stakes are low and no wrong number touches a tenant or a lender.
Draw the line where output touches the firm's record or a counterparty: rent rolls, LOIs, broker updates, lease abstractions, and anything sent outbound. That work belongs on a governed desk.
Utility belt vs. governed desk — which is for whom?
- Utility belt (general chat tools): for individuals doing low-stakes drafting and research. Cheap, instant, no oversight built in.
- Governed desk (system of record + approvals): for firms where actions touch shared records and counterparties. Slower to stand up, but every action is proposed, approved, and logged.
Candidly: if you're a one-person shop and nothing you produce is binding, the belt is enough. If you have a rent roll, brokers, and lenders who expect the number to be right, you need the desk.
FAQ
FAQ
Q: Do I need to throw out ChatGPT and my other AI tools? No. General tools are fine for low-stakes, single-person drafting and research. The point is to stop them from touching your firm's records or sending anything to a counterparty without a person approving it first.
Q: What does "propose, never commit" mean in practice? The system stages an action — a follow-up, an alert, a record update — and a human approves it before anything happens. Nothing is sent, changed, or committed autonomously, and every approval is logged.
Q: Isn't a governed desk just slower? It adds an approval step, which is the point. Generation is already fast; the expensive mistakes come from unapproved sends and unchecked numbers. The desk keeps the drafting speed and removes the silent commits.
Q: Can a governed desk still do prospecting and outbound? Yes. It can gather prospect data freely and draft outbound at scale — but sending waits for a human to sign off, so wrong names and wrong numbers don't reach the market.