For the most ambitious lenders in real estate
- Underwrite the property from the loan request package, checked against itself and against the market.
- Test the property on the borrower’s numbers and on market numbers: LTV, DSCR, debt yield.
- Every red flag on one page. Every input sourced.
- Banks
- Debt funds
- Private lenders
- Credit analysts
- Underwriters
The loans coming due in your market, found for your desk
Signals is the Realiste AI agent that finds deals and borrowers. Tell us the market and the loan sizes you write, and we send you the list: every securitized loan maturing there, with the property address, the balance and the maturity date on each row.
- Your marketA state, a metro or a county, and the property types you lend on.
- Your windowLoans maturing in the next 6, 12, 24 or 36 months.
- Your listOne row per property: address, balance and maturity date, ready for your CRM.
Name your market and window in the message box. We reply by email with the list.
You send the borrower package as it came
- Offering memorandum
- Rent roll
- T-12
- The borrower’s model
Property documents only. Leave out personal financial statements and anything with an SSN.
You get back
- The package checked against itselfRent roll against T‑12 against the offering memorandum and the model. Every gap with the amount and the page it sits on.
- The package checked against the marketRents, vacancy, expenses, cap rate and sale comps around the property, each with its source and how far the borrower’s figure sits from it.
- Coverage tested twiceOn the borrower’s numbers and on market numbers: LTV, DSCR and debt yield at your limits, in the workbook. The loan amount stays with your committee.
- Red flags on one pageWhat the credit committee should see first, and the questions to put to the borrower.
- Excel with live formulasThe sizing and the cash flow in a workbook your analysts can change and recalculate.
- A source on every inputEach input names the document and page, or the record and date, it came from; each result shows its formula.
1017 Jefferson Ave, Miami Beach
A 1925 multifamily building, underwritten from property records and market data, with no borrower package. The same check runs on the borrower’s numbers when the package is there.
- Year 1 NOI
- $171,703
- Going-in cap rate
- 4.29%at the $4,000,000 reference price, the last recorded sale
- Units on record
- 21 or 26two records disagree; named, not averaged
The workbook carries LTV, DSCR and debt yield at a 65% LTV and 1.25x DSCR test, as live formulas your analyst can change.
Red flags the run raised
- Legal unit count unresolved: one record shows 21 units, another 26.
- A lien flag on the property, with no type, amount or date on record.
- A legally nonconforming building under its zoning, so rebuild rights need a city letter.
- The building recertification required of a 1925 building is not verified.
Run on a real address as a benchmark, not a client deal. Illustrative of the check, not a recommendation to lend.
Your stack, any stack. Our engine in the middle. Answers land where you work.
Everything you have now stays yours. It runs through the Realiste engine for real estate, gets enriched with our sources, and the conclusions come back to wherever you read them. Set up once, it runs on its own: no moving files by hand, no clicking around.
What you already have
- Your inbox
- Your data rooms
- Your CRM
- Your Excel models
- The territory you set
The Realiste engine for real estate
- Reads it
- Enriches it with 2,100+ connected sources, 1,881 already returning data
- Draws the conclusions
Lands where you like
- Telegram
- Your CRM or model
Your list, enriched
Your loan book’s property list comes back with the maturities we hold, owner, permits and dated events (tax delinquency, foreclosure filings), matched by our team at the start. What we could not match is named, not guessed.
- Yours alone: never shown to another client
- NDA first
- Your sign-off on everything that goes out
Your AI assistant can ask the engine directly, too.
Your workflows, on our engine
Each one starts where your work already starts and ends where you read it. Two run for every team: the inbox and the radar. The others are yours.
A loan request lands in your inbox
- Trigger
- A broker sends the borrower’s OM, T‑12 and rent roll to your loan request inbox, and it goes on to the engine by itself.
- Engine
- Spreads it, ties it out, adds the owner of record, liens, court and tax records and sales comps, and tests coverage and debt yield at your rates.
- Lands
- Back in the same thread after our team checks it: your spread filled, the credit memo draft and the flags with their pages. The loan amount stays with your committee.
The radar on your footprint
- Trigger
- In your footprint, a loan comes due within 24 months, coverage weakens or a foreclosure is filed.
- Engine
- Adds the owner behind the LLC and the reported coverage and occupancy, each with its date.
- Lands
- A weekly digest by e-mail, Telegram or WhatsApp: the borrowers who will need a lender.
The weekly book watch
- Trigger
- Your collateral list, matched to our property records once, at setup.
- Engine
- Checks every property each week: taxes delinquent, a new mortgage recorded, a court case filed, code violations, sales nearby.
- Lands
- A weekly e-mail of what changed, loan by loan, with the next step; DSCR re-tested on the latest T‑12 for a loan that slips.
A construction loan request
- Trigger
- A construction or BTR request arrives with a site address.
- Engine
- Reads what the zoning allows, prices the schemes that fit and checks the site: owner of record, open permits, liens.
- Lands
- A live Excel and a memo in your inbox, every zoning figure with its source; a feasibility study takes about half an hour.
More accurate, faster, lower cost per deal
More accurate
The package checked against itself and against what is on record; a gap is named with its amount and page, never filled in.
- Every input shows its source and date; every result shows its formula.
- What we could not find is named as missing, never filled in.
- The same comps request returns the same set. The same feasibility request returns the same study for 30 days.
Faster
A property underwriting from the loan request package in about an hour, plus a due-diligence checklist where every item points to its document and page.
| Property passport | about 4 minutes |
|---|---|
| Comps | about 20 minutes |
| Feasibility study | about 30 minutes |
| Up to 10 properties valued at once | about 30 minutes |
| Value opinion | about 40 minutes |
| Underwriting with its live Excel model | about 1 hour |
Lower cost per deal
- One request returns the finished package: the memo, the live Excel, and the deck on request.
- An underwriting carries about 1,000 data points, each marked as found with its source, computed by formula, or missing.
- Your analysts check and decide instead of collecting.
BOV at scale
Up to 100 addresses, priced in one go
Paste a list of up to 100 addresses and get one value range per property, each with its comparable sales and a confidence mark, as a table, a workbook and a deck. A list of ten comes back in about half an hour. The list size is set for each client, and we raise it for yours.
Mark up to 100 collateral properties in one go, each with a value range and its comps.
A value range with its evidence, not an appraisal. Realiste is not an appraiser.
Up to 100 rows, one per address, as a table, a workbook and a deck.
Your processes, on AI
Loan underwriting support from intake to committee. Realiste sits beside your loan origination system; it does not replace it. We take one lending process your team runs today and run it in your own templates; your team reviews and signs.
Screening a loan request
- Comes in
- The OM, T‑12, rent roll and leases, as they reached your inbox or data room.
- Comes back
- Your spread, filled. The rent roll tied to the T‑12, the OM checked line by line, each gap with its amount and page. Leases abstracted where they are in the package (text PDFs; scanned pages are flagged, not read).
- Template
- Your spread, your tabs, your formulas.
- Replaces
- Re-keying the package and the tie-out by hand.
The credit memo draft
- Comes in
- The screened package, the amount requested, your rates and your memo template.
- Comes back
- A memo draft: collateral, owner of record against the borrower entity, liens and court records on the collateral, market rents and sales comps, NOI three ways, coverage and debt yield across rates, findings and questions for the borrower.
- Template
- Yours, with the live Excel behind every number.
- Replaces
- An analyst’s day assembling the memo. The loan amount and the decision stay with your committee.
Collateral page and pre-funding checks
- Comes in
- An address, or the properties on a borrower’s schedule of real estate owned.
- Comes back
- Parcel and owner of record, sales, recorded mortgages and liens, court and tax records, zoning and nearby sales. Each category marked found, none found or no source here.
- Template
- One page, or the fields dropped into your credit file.
- Replaces
- Looking up each record by hand. Not a title commitment: a fast read before you order one.
Construction and BTR loans
- Comes in
- The site address or parcel, and the borrower’s planned project.
- Comes back
- What the zoning allows (district, floor area ratio, height, maximum buildable area), the schemes that fit with cost and return for each in a live Excel, and site diligence: owner of record, open permits, liens.
- Template
- Your construction credit file, every zoning figure with its source.
- Replaces
- Weeks of waiting on a zoning read before the appraisal is ordered.
Your book and the watch list
- Comes in
- Your collateral list, matched to our property records once, at setup.
- Comes back
- Each week, what changed at each property, with its date: taxes delinquent, a foreclosure or court case filed, code violations, a new mortgage recorded, a sale nearby. For a loan that slips, DSCR re-tested on the latest T‑12.
- Template
- Your watch-list format.
- Replaces
- Finding a tax default or a new lien months after it was recorded.
New business
- Comes in
- Your footprint: states or metros, property types, loan size.
- Comes back
- Loans on record in your footprint that mature within 24 months, each with its maturity month, reported coverage and occupancy with their dates, and the owner behind the LLC.
- Template
- A sheet your CRM imports.
- Replaces
- Market reports that never name a borrower.
The products behind it
Each process chains these products. On top of them sits one assistant and 330+ written workflows, 34 of them for lenders; we write yours.
| Product | You give | You get back |
|---|---|---|
| Underwriting | OM, T‑12, rent roll | A memo and live Excel: NOI three ways, DSCR and debt yield, stress grids |
| Due diligence | The data room | Eight stages, each item ok, issue or missing, with its page |
| Property passport | An address | One dated page per collateral: owner, sales, recorded debt, liens; gaps named |
| Comps and market rent | An address, a rent roll | Sales of the same type nearby, and a rent check |
| Signals | Your counties, your filters | A weekly list we pull for you: loans coming due, distress, new owners |
| Feasibility | An address or parcel | For construction loans: the zoning read and a live Excel per scheme |
| Excel by prompts | Your workbook, up to 40 MB | The same file, changed by typing; every change shown first |
Where each team starts
| Your team | Start with | Why there |
|---|---|---|
| Bank CRE desk | Screening and the credit memo in your spread and memo format | The volume is in requests, and every memo follows one fixed order. |
| Debt fund, bridge lender | Maturities in your markets, then screening the requests they bring | Your next borrowers are owners whose loans come due, and speed to a term sheet wins. |
| Construction and BTR lender | The site and project check | The risk sits in the site and the exit, before the first draw. |
| Private or family-office lender | The collateral page and pre-funding checks | A small team, quick closings, and the collateral is the whole story. |
| Loan broker, debt advisor | A property underwriting from the borrower’s package, spread and checked before it goes to lenders | You win by reaching the borrower first and sending a clean file. |
One process, live on your loans in 72 hours
Counted from access to your stack, after the NDA, with your team, on live loans.
Day 1
Map
We sit with the person who does the work and write the process down: what comes in, what gets checked, your template, who signs.
From you: An hour with the process owner
Day 2
Build
Our team connects it to your stack (where the work arrives, your models, your CRM or warehouse) and to your template, then runs it on one live case.
From you: Your template and one live case
Day 3
Run and measure
Your team runs it on live cases. Together we time each one and check every figure against its source.
From you: Two or three more cases
Then
An annual agreement
The process keeps running and we tune it to your team. The next process is mapped, built and measured the same way.
From you: The next process you pick
Questions teams ask first
- How do lenders screen commercial real estate loans for default risk faster?
- They read the borrower’s package against itself before the credit team starts: the rent roll against the T‑12, both against the offering memo’s NOI, then debt coverage and debt yield at two rates. Realiste runs that tie-out as one run and returns each difference with the page and cell behind it; on one 50-unit loan package it showed the memo’s in-place NOI 32% above the T‑12, while the rent roll and the T‑12 agreed with each other. It screens the file and does not score default or make the credit call.
- Do we have to change how we work?
- No. Your stack stays yours. What arrives in your inbox, your data rooms and the territory you set runs through our engine, and the answers land where you work: e-mail, WhatsApp, Telegram or your CRM, as Excel, PowerPoint, PDF and Word. Nothing new to log into.
- What do you bring, and what do we keep?
- 2,100+ connected sources (parcels and owners of record, recorded sales and loans, permits, zoning and rents) and a property base of about 6 million properties. Your files serve your work, are never shown to another client and are not used to train models. An NDA comes before any access.
- Does Realiste make the credit decision?
- No. Realiste gives no credit opinion and no appraisal of its own, and the memo states no loan amount. The coverage and debt yield math sits in the workbook; the decision to lend stays with your committee.
- How long until it runs on our own work?
- 72 hours from access, after the NDA. Day 1 we map the process with you, day 2 we build it into your setup and run one live case, day 3 your team runs it and we measure it together.
Your borrower’s files stay yours
Each firm’s data is kept isolated and never used to train models. Nothing you send is published or shared with other customers, and every run records who ran what, when, and on which sources.
How we handle your data →Decision support, not a decision
Realiste gives no credit opinion and no appraisal of its own. It shows the numbers, where they came from and where they disagree. The decision to lend is yours.