An owner who asks a broker what a property is worth can have a sourced value opinion draft back in about an hour, with the broker still signing it. In one measured run on the Titleman engine, a single typed line, "Value opinion for" and a street address, returned a memo and a live Excel workbook in 62 minutes, built from 257 records and 73 inputs that each carry their source. The run also caught that the address is one 750 SF bay, not a parcel: what can be sold is the whole 9,772 SF strip. This page walks the workflow step by step, including the four things the draft said it could not establish.
Written by the Titleman team with AI assistance (Claude). Reviewed by the Titleman review desk. The run was made on a Titleman QA account on October 14, 2026, and every count below was recounted from the finished run. It is a measured sample on a real retail strip in Miami with the address withheld. It is not a client's deal and not an appraisal.
The workflow: an owner emails "what would it trade for?"
The owner wants a number, the broker wants to answer fast, and the draft behind the answer is usually rebuilt by hand for each owner who asks.
| Step | By hand | In the run on this page |
|---|---|---|
| 1. Question in | Read the email, find the property | One line typed into the workspace: "Value opinion for" and the address. No files, no rent roll |
| 2. Read the property | Pull the parcel record, owners, taxes, zoning, permits, flood zone | 257 records read |
| 3. Find the evidence | Search sales and asking rents, rebuild a table | 6 nearby sales found, 3 used; asking rents and cap-rate surveys read |
| 4. Price it | Build the workbook | 73 inputs, each with its source; income and sales methods, land as a check |
| 5. Write it up | Draft the memo | A memo in PDF and Word and a live workbook |
| 6. Review and send | The broker reads, edits and signs | Unchanged. The broker of record still reviews and signs |
What the engine returned
The finished run delivered four things, in one answer:
- A memo in PDF and Word: the facts of the property, a table of red flags with dates, the price factors, and a highest-and-best-use read.
- A live workbook. Change the rent, vacancy, cost or cap-rate inputs and each method moves, so the broker can see how the number depends on each assumption.
- A sales table with every sale the run considered, which of them it used, and why it set the others aside.
- A list of what it could not establish, with the person who holds the answer.
The memo does not state a price. It gives the facts, what the price depends on, and how each method is worked out, so the broker chooses the number and can defend it.
What the run caught: the address was not the asset
The request named one address. The records showed that address is a single 750 SF bay in a one-story retail row, held in one record with a 9,772 SF building built in 1982 on a 28,464 SF lot. Business licenses were active at nine of the bay addresses. The bay could not be sold on its own, so the factors and methods describe the whole strip.
That is the kind of catch a quick hand pull often misses. An owner who asked about the bay would have been priced as if the bay were the property.
What it said it could not establish
A draft that hides its gaps is worse than no draft. This one listed four:
- The rent roll, the leases and the operating statement. None was available, so the in-place income is unknown. The owner or the listing broker holds them.
- Whether the bay is empty. It was offered for lease while a business license at the same address was still active.
- The buildable area under the zoning district. The record held no floor-area or density standard, and a zoning verification letter would close it.
- An outside sale listing. Its page blocked the read, so the date, broker, size and income were not confirmed.
The sales evidence was also weak. Both recent sales of a usable kind were coded as non-market transactions, and the one qualified sale was older and less than half the size. The draft says so, and it ranks the evidence accordingly.
Time: what is measured and what is not
The measured number is 62 minutes, from the typed line to the delivered answer, on October 14, 2026. We did not time an analyst doing the same job by hand, so this page does not state a saving as a number, and it would be wrong to read one into the table above. The useful comparison is your own: time the last owner request your team answered, from the email to the sent reply, and set it beside 62 minutes. For how single-property runs are distributed, see how long a broker opinion of value takes.
What the broker sends the client
The broker sends a reviewed value opinion, not the raw draft:
- A cover note in the broker's own words, with the number the broker has chosen and the reason for it.
- The memo, edited by the broker, with the list of open items turned into questions for the owner (for example, "please send the current rent roll").
- The workbook, if the broker wants the owner to test the assumptions.
- The broker's signature. The draft is for the licensed broker of record to review and sign. Titleman does not sign it and does not send it to the client.
What the 72 hours are
The run above took 62 minutes. The 72 hours is something else: the time our team needs to wire one workflow of yours into the engine. In the first case, an owner's email goes through the engine and the draft comes back in the same thread, with your CRM, WhatsApp or Telegram as other places the answer can land. You do not move anything. Your stack stays, the engine reads from 2,100+ sources, and the answer arrives where you already work. The run on this page was typed into the workspace; we have not run a live broker inbox for it and do not claim one.
Frequently asked questions
How long does it take AI to produce a value opinion draft for an owner who asks what a property is worth?
In the run on this page, 62 minutes from one typed line to a delivered memo and a live Excel workbook. That is one measured run on a 1982 retail strip in Miami, not a promise for every property. Single-property runs vary, and a run that finds a disagreement in the records takes longer because it resolves it before writing anything down.
Can a broker send the AI draft straight to the owner?
No. The output is a draft for the licensed broker of record to review and sign. The run on this page listed four things it could not establish, and the broker closes those, or says plainly that they are open, before anything goes to the client.
Is a value opinion draft an appraisal?
No. It is an opinion of value prepared for a broker's work. Titleman holds no appraisal license, and in the United States appraisal is licensed work. For the difference in practice, see broker opinion of value vs appraisal.
What does the 72 hours mean?
The 72 hours is the start-up job, not the run time. Our team wires one workflow of the broker's own, for example an owner's email that comes back with the draft in the same thread, within 72 hours of getting access. The run itself took 62 minutes.
Where to go next
Name one workflow in your inbox, for example an owner asking what a property is worth, and we show you the same draft wired into it. See Titleman for brokers, the broker opinion of value software page, and anatomy of one BOV-at-scale run for what sits under a run.