What is an AI-assisted CRE debt broker vetting process? It is a structured way to evaluate a commercial mortgage broker before you sign an engagement letter, using AI to reconcile the broker's claims against public origination data, extract the terms buried in the fee agreement, and generate reference questions that expose a thin placement record. Most borrowers vet the loan carefully and the broker barely at all, which is backwards: the broker decides which lenders ever see your deal. For how AI fits across the financing stack, see our guide to AI CRE finance and capital markets.
Key Takeaways
- A debt broker's value is lender access and execution certainty, not rate shopping, so vet the relationships and the closing record rather than the pitch deck.
- AI is best used here to reconcile claims against evidence: origination rankings, closed-deal lists, county records, and the broker's own marketing all in one comparison.
- The fee agreement matters more than the fee. Exclusivity windows, tail periods, and fee triggers on loans you source yourself decide what you actually owe.
- Match the broker to your capital source. CBRE reported alternative lenders at 53% of Q1 2026 non-agency loan closings, so agency-only shops may not reach your best execution.
- Ask for the deals that did not close. A broker who cannot describe a failed placement honestly has either no volume or no candor.
When You Actually Need a Debt Broker
You need a debt broker when your deal does not fit the lender you already know. If your fourth loan is the same asset type in the same market at the same leverage as the three you closed with your regional bank, a broker adds a fee and a layer. If you are financing an asset type you have never owned, entering a new market, chasing higher leverage, or working against a hard closing deadline, a broker earns the fee by knowing which twenty lenders to call and which forty to skip.
Fees are quoted as a percentage of the loan amount, commonly in the 0.5% to 1% range on institutional-size deals and higher on smaller loans. Nothing is standard and everything is negotiable. The honest test is whether the broker's access produces an execution you could not have reached yourself.
Market structure argues for breadth. CBRE reported alternative lenders as the largest non-agency lender group in Q1 2026 at 53% of loan closings, ahead of banks at 22% and life companies at 17%. If your broker's entire book is agency business through Fannie Mae and Freddie Mac, that is fine for stabilized multifamily and a real gap for a transitional office or industrial deal that needs a debt fund.
The Five Things That Separate Debt Brokers
Five variables actually distinguish one debt broker from another, and none of them appear on a pitch deck in usable form. Rank every candidate on these before anything else.
- Asset class placement record: Loans closed in your property type in the last 24 months, not career totals across every asset class.
- Capital source coverage: Which lender categories they actually close with: banks, life companies, debt funds, CMBS conduits, Fannie Mae, Freddie Mac, and HUD or FHA through Ginnie Mae.
- Deal size band: A shop that closes $50 million loans will not staff your $4 million deal well, and a small-balance shop cannot reach institutional capital.
- Execution rate: The percentage of engaged deals that reach closing, and what happened to the ones that did not.
- Who does the work: Whether the producer who pitched you will run your file or hand it to an analyst you have not met.
The last one is the most common disappointment and the easiest to prevent: get the name of the person who will assemble your loan package into the engagement letter.
Building the Vetting File With AI
AI earns its place here by reconciling sources, not by having opinions. Collect the broker's pitch deck, closed-deal list, team bios, and website copy, upload them to ChatGPT, Claude, or Gemini in one project, and ask a specific reconciliation question.
A prompt that works: Extract every closed transaction claimed in these materials into a table with date, property type, market, loan amount, lender name, and lender category. Flag any transaction where the lender is unnamed, the date is missing, or the amount is a range. Then summarize the distribution by property type, lender category, and loan size band. The gaps are the finding. A closed-deal list where 40% of entries have no named lender is telling you those were participations or referrals rather than placements.
Then test the claims against a public source. The Mortgage Bankers Association publishes annual commercial and multifamily originations rankings covering 129 originators across more than 140 categories, broken out by investor group, property type, financing structure, and originating office location. That last cut matters: it tells you whether the firm's volume comes from the office that will handle your deal or from another region. Eastdil Secured, Walker & Dunlop, Berkadia, JLL, CBRE, and Newmark appear across those league tables, and a broker claiming peer-level volume who appears nowhere is worth a direct question.
Use AI on your side of the ledger too. Run your deal through an underwriting pass first so you know your DSCR, debt yield, and LTV cold, using the mechanics in AI DSCR analysis and the lender-category context in AI debt fund analysis for CRE lending opportunities. A borrower who knows their own numbers is harder to steer toward whichever lender pays the broker best.
The Reference Call Script AI Should Write for You
Generic reference calls produce generic answers. Have AI generate questions from the specific gaps it found in the vetting file, then take those to the borrower references the broker provides, asking for questions that cannot be answered with a yes. The ones that consistently produce signal:
- What changed between the term sheet and the closing? Rate, proceeds, and reserves almost always move. You want to hear how the broker handled it, not that it never happened.
- How many lenders quoted, versus how many the broker said he would approach? A gap between twenty-five approached and three quotes is a marketing problem or a deal problem, and you want to know which.
- Who did you deal with after the engagement letter was signed? This surfaces the producer-to-analyst handoff without accusing anyone.
- Would you use them again on a deal with a hard deadline? The qualifier separates polite references from real ones.
Also ask the broker directly for a deal that did not close and why. One with real volume has several and will describe one plainly, including their own missteps. A broker who cannot produce one is either too new or is managing you. The AI Consulting Network builds these vetting workflows as reusable templates rather than one-off exercises.
Reading the Fee Agreement Before You Sign
The engagement letter, not the fee percentage, determines what a broker relationship costs you. Extract these terms with AI before you sign, because they are usually scattered across dense paragraphs and easy to skim past.
- Exclusivity period: How long you are barred from talking to lenders directly, and whether it auto-renews.
- Tail period: Whether a fee is owed on a loan closing after the engagement ends with a lender the broker introduced. Commonly six to twelve months and legitimate, but the introduced-lender list must be defined in writing.
- Fee trigger: Whether the fee is earned on closing and funding or merely on a signed term sheet. Insist on closing and funding.
- Self-sourced carve-out: Whether you owe a fee for closing with a lender you brought. Name your existing relationships in an exhibit before signing.
- Retainer and expenses: Whether the retainer credits against the fee, and which third-party costs you carry.
- Dual compensation: Whether the broker also collects from the lender. Ask in writing, get the answer in writing.
Once quotes arrive, the same extraction discipline applies to the term sheets. Our guide to AI for CRE lender term sheet comparison covers normalizing competing quotes to an all-in cost of capital, and AI loan comparison tools for CRE covers the modeling. Vetting the broker and vetting the quotes are different jobs, and doing the first well makes the second shorter.
A Repeatable Broker Scorecard
Turn the vetting file into a scorecard you reuse on every financing, because the value compounds only if it is repeatable. Score each candidate 1 to 5 on the five variables above plus fee agreement fairness, weighting capital source coverage and execution rate most heavily since those are what you cannot replicate yourself.
Have AI populate an evidence column with the document or league table entry behind each score, and leave any score without evidence blank rather than guessing. A blank cell is a question for the next call. The same evidence-first discipline shows up in adjacent screening work, whether you are vetting CRE crowdfunding deals as a passive investor or sourcing and screening distressed CRE debt.
Two limits are worth stating plainly. AI cannot verify a relationship, and relationships are most of what you are buying. It also cannot tell you whether a broker answers the phone at 7pm on a Friday when a rate cap quote expires Monday. Those you learn from references and the first deal. Avi Hacker, J.D. at The AI Consulting Network works with CRE owners and sponsors on turning this into a standing process.
Frequently Asked Questions
Q: How much does a CRE debt broker charge?
A: Fees are quoted as a percentage of the loan amount, commonly in the 0.5% to 1% range on institutional-size deals and higher on smaller loans. Nothing is standard and the fee is negotiable. What matters more is when the fee is earned, whether a tail period applies, and whether the broker also collects from the lender.
Q: Can AI actually tell me if a debt broker is good?
A: No, and that is the wrong job for it. AI reconciles the broker's claims against public origination data, extracts the real terms from the fee agreement, and generates sharper reference questions. It cannot verify lender relationships or judgment under deadline pressure, which is most of what you are hiring.
Q: Should I use a debt broker or go directly to lenders?
A: Go direct when your deal matches a lender you have already closed with at similar leverage and asset type. Use a broker for a new market or asset class, higher leverage, a hard deadline, or access to debt funds and life companies you do not already bank with.
Q: What is a tail period in a debt broker engagement letter?
A: A tail period means you still owe a fee if you close after the engagement ends with a lender the broker introduced, typically for six to twelve months. It is legitimate protection against being cut out, but insist the introduced-lender list be defined in writing.