Titleman Signals finds the properties in your territory where a deal is more likely to happen, from a loan coming due to a new permit, and puts the AI to work on that list against your brief. On Titleman test and staff accounts, 7 delivered runs between 15 and 22 October 2026 took a median of about 31 minutes.
Published by Titleman, which makes the product described. Updated 22 October 2026.
Specification at a glance
| Item | Specification |
|---|---|
| Job | Finds the properties where a deal can happen next and works the list against your brief |
| Signals | 48 in 7 groups |
| Territory | City, county, corridor, address, or an area you draw |
| Outputs | Summary, ranked table, list of what was not checked, next steps, Excel workbook |
| Median time | About 31 minutes (7 delivered runs) |
| Fastest and slowest | About 22 and about 50 minutes |
| List worked | 9 to 200 properties in those runs |
| Accuracy | No error rate measured on real client work |
Inputs
You choose the place, the signals and a brief. The count of matching properties updates as you choose. The brief is a few lines on what the AI should find in the list: owners to call, the lender and the loan terms, permits filed this year. An empty brief still runs, and the AI then ranks the list by how strong and how near each property's signals are and checks the strongest. You can attach your own files, such as a list of owners you already know or your buy box.
| Group | Signals |
|---|---|
| Debt | 10, from maturing and prepay window open to value below loan |
| Deals and market | 3: just sold, serial buyer, serial seller |
| Owner | 6, including long hold, absentee owner and portfolio owner |
| Distress | 8, including tax delinquent, foreclosure filed and code violations |
| Income and tenants | 6, including occupancy down and anchor lease expiring |
| Development sites | 12, including vacant land, teardown and new construction permit |
| Regulation and affordable housing | 3: LIHTC year 15, subsidy expiring, mandated capex |
Outputs
- A summary: what you asked, how the list was narrowed, how many properties were worked, and the strongest few.
- A ranked table: each property, why it fits, the data found for it with where each fact comes from, and what is still unknown.
- Not checked on this run: the properties of the list the run did not reach.
- Next steps in a few lines, and the ranked table as an Excel workbook.
What the AI looked for and did not find is marked as not found.
Measured time
Time is counted from the start of the run to the delivered answer. The figures come from 7 delivered Signals runs on Titleman test and staff accounts, created between 15 and 22 October 2026: median about 31 minutes, fastest about 22, slowest about 50. The list worked ranged from 9 to 200 properties, and time did not follow list size: a run on a 9-property list took about 42 minutes, and a run on a 200-property list took about 22. Of the 11 Signals runs on those accounts in the period, 7 were delivered, 2 ended within a minute with a question back about what kind of property it was, and 2 ended in an error.
Limits
- A list, not a building. Signals works a list. It does not value or underwrite one property.
- Coverage varies by county. What your territory holds is counted before the first run.
- Large territories. A run works the strongest properties of the list. The rest are named under Not checked on this run.
- Open to selected accounts. If yours is not open, the page says so and nothing else happens.
- Sample size. Seven runs is a small sample, and the times will move.
What it does not replace
Signals does not replace your judgment about who to call, a valuation, an underwriting or a visit. A signal is a reason to look at a property, with its date and the record behind it. It does not say the owner will sell.
How these figures were measured
Source: run records of the Titleman AI engine, read on 22 October 2026. Which runs: all Signals runs on Titleman test and staff accounts created between 15 and 22 October 2026 (11 runs: 7 delivered, 2 ended with a question back, 2 ended in an error). Runs on client accounts are left out. The median is rounded up to the whole minute. List sizes are the number of properties each run states it worked, as written in the run's own summary.
Frequently asked questions
How long does a run take? A median of about 31 minutes in 7 delivered runs, between about 22 and about 50.
Which signals? 48 in 7 groups.
How big a list? 9 to 200 properties worked in those runs. The rest are listed as not checked.
What comes back? A summary, a ranked table, what was not checked, next steps and an Excel workbook.
Does it value a property? No.
To get the list for your own territory, request a demo. The product page is Signals.