Blog
AI Prompts for Real Estate Investors (2026)
By Jaken Finance Group · Principal, Jaken Finance Group
Worked AI prompts for flips, rentals, and wholesale screens — plus the ARV, MAO, and hold-cost checks that keep ChatGPT from blowing up your offer.
AI prompts for real estate investors only work when you feed the model real inputs and verify every number before you offer. A one-line question — “Is this a good deal?” — gives you a confident paragraph built on nothing. A structured prompt gives you a worksheet you can check against sold comps, a contractor bid, and a lender file.
Most “60 prompts” listicles online are vendor ads. They give you questions like “What is the APN?” or “Will I lose money on this flip?” without context, constraints, or a verification step. You paste them, get a polished answer, and mistake speed for accuracy. The competitive edge is not having more prompts — it is running the same underwrite workflow every time and checking the output against primary sources.
This guide is tool-agnostic. Paste these templates into ChatGPT, Claude, Gemini, or a property-data chat tied to a subject property. Then run the verification steps at the end of every section. For the math behind the prompts, pair this with how to calculate ARV and the 70% rule.
Who this playbook is for
- New fix-and-flip investors building their first comp and MAO discipline
- Wholesalers who need buyer-ready packets without inventing ARV
- Buy-and-hold and BRRRR operators screening rent and DSCR before they tie up capital
- Experienced sponsors who want faster first drafts — not a substitute for appraisal, title, or bids
You will get 20 copy-paste templates grouped by workflow: find → underwrite → rehab → rent or sell → finance → verify.
How do you write an AI prompt that survives a lender file?
A lender-ready prompt has five parts: role, inputs, constraints, output format, and a verification step. Skip any one and the model fills the gap with a guess.
| Part | What to include | Example |
|---|---|---|
| Role | Who the model is helping | ”You are a fix-and-flip underwriter assistant.” |
| Inputs | Facts you paste — address, beds/baths, sq ft, your comp list | Subject + 3–5 sold comps with dates and prices |
| Constraints | Rules the model must follow | ”Sold only. Last 180 days. Same neighborhood. No list prices.” |
| Output | Table + assumptions + unknowns | ARV table, MAO line, red-flag list |
| Verify | What you will check outside the chat | ”Flag every number I must confirm in MLS or with a contractor.” |
Bad prompt: “Is this a good flip?”
Good prompt: “You are a flip underwriter. Subject: 3/2, 1,420 sq ft, built 1978, [CITY]. I will paste five sold comps. Build ARV from median adjusted sold price. Calculate MAO at 70% minus $48,000 rehab minus $12,000 hold. List assumptions and unknowns. Do not use list prices or tax assessed value.”
Key numbers to bake into every underwrite prompt
- Comps: 3–5 sold; last 90–180 days; similar size and finish level
- MAO screen: (ARV × 0.70) − rehab − estimated hold
- Lender ARV cap: often 70%–75% of appraised after-repair value
- Hold stack: interest-only carry plus taxes, insurance, utilities — see holding costs
Which AI tool for which job?
| Task | Best fit | Why |
|---|---|---|
| Long worksheets, tables, MAO math | ChatGPT, Claude, or Gemini | Strong at structured output when you paste inputs |
| Property-specific Q&A on one record | Your data vendor’s property chat | Pulls APN, last sale, liens from their feed — still verify at county |
| Marketing copy, seller letters | Any general model | Low stakes — edit before sending |
| Skip trace, owner phone, lien clearance | Not AI | Use title, county, licensed skip-trace |
Per the NAR 2025 Technology Survey, ChatGPT was the most cited AI tool at 58% among agents who use AI, followed by Gemini (20%) and Copilot (15%). Tool choice matters less than prompt structure.
Save your prompt library
Create one note per strategy (flip, rental, wholesale) with your bracket fields filled in: [CITY], [BEDS], [REHAB_CAP]. Update quarterly when your hold assumptions or lender caps change. Same prompts, fresh numbers — that is how teams stay consistent.
What does AI get wrong on investor deals?
AI is fast at structure and slow at truth. The NAR 2025 Technology Survey (July 2025, 1,241 respondents) found 46% of agents use AI-generated content, but 46% said AI had no noticeable impact on their business. The gap is not access — it is verification.
| AI output | What actually decides the number |
|---|---|
| ”ARV is $340,000” | Your sold comps + appraiser bracket |
| ”Rent is $2,100/mo” | Rent comps, lease, or Form 1007 |
| ”Rehab is $35,000” | Contractor bid or line-item scope |
| ”Owner is motivated” | Public records only — label as hypothesis |
| ”No liens” | Title search or county recorder |
| ”Monthly payment at 7%“ | Your actual loan quote — Jaken Finance Group fix-and-flip runs 8.99%–13.5% on qualified files |
| ”Net profit $45,000” | Full P&L with sell costs and hold — gross spread is not net |
Never paste skip-trace phone numbers, SSNs, or full owner dossiers into a public model. Use AI on property and deal math, not on harvesting contact data.
What we did not publish from the source listicle
Roundups that promise “instant profit” from a property chat often mix data lookup (“What county is this in?”) with underwriting (“What should I offer?”) without separating them. Lookup prompts are fine if your vendor has the field. Underwrite prompts fail when the model invent comps or treats list price as ARV. We also omitted vendor-specific product pitches and unsourced rules like a fixed “15% ROI certainty premium” by zip code — no primary methodology supports that as a universal wholesale screen.
The gross-vs-net trap in AI flip answers
If you ask “Will I make money on this flip?” the model will often subtract purchase from ARV and call the difference profit. That is gross spread, not net. Rehab, hold, buy-side and sell-side closing, and commission come out of that spread. When you prompt for MAO, explicitly require sell costs and hold in the output so the model does not hand you a fake win.
Find and screen: four prompts before you underwrite
Use these when a lead first lands. They organize facts you already have — they do not replace a property-data subscription.
1 — Property fact sheet (from your pasted data)
When: You pulled tax, listing, or county records and want a clean one-pager.
You are a real estate acquisition analyst. I will paste raw property data below.
Build a one-page fact sheet with: address, APN, year built, beds/baths/sq ft, lot size,
last sale date and price, tax assessed value (labeled "not market value"), HOA if known,
and zoning if stated in my paste. List every field you could not confirm from my paste as
"UNKNOWN — verify." Do not invent owner contact information.Verify: Cross-check APN, last sale, and tax against county recorder or your data vendor.
Good output looks like: A clean table with a “Source: user paste” column and a short “UNKNOWN” list — not a narrative that fills gaps with plausible guesses.
2 — Red-flag checklist
When: Before you spend an hour on comps.
You are a conservative flip buyer. Subject property facts:
[PASTE ADDRESS, YEAR BUILT, BEDS/BATHS, SQ FT, PRICE, HOA, FLOOD ZONE]
List red flags in three buckets: (A) walk-away, (B) price the risk in rehab or hold,
(C) verify with inspection/title. Cover: foundation/1970s systems, polybutylene, aluminum
wiring, flood/elevation, HOA rental caps, non-arm's-length last sale, long vacancy,
environmental, access/easement, illegal conversions. Mark each item VERIFY or BID REQUIRED.
Do not estimate repair dollars without a "bid required" label.Verify: Inspection, title commitment, FEMA flood map, HOA docs.
3 — Owner-motivation hypotheses (labeled as guesses)
When: You have public record signals — equity, years owned, absentee, pre-foreclosure flag from your vendor.
You are an acquisition analyst. I will paste public-record owner and mortgage facts only.
List up to five HYPOTHESES for why the owner might sell, each tagged HYPOTHESIS.
For each, cite the public fact that triggered it. Do not state motivation as fact.
Do not provide skip-trace or contact steps. End with three questions I should ask the seller
on the first call.Verify: Seller conversation. Public records do not prove distress.
4 — Walk-away criteria
When: You want the model to kill bad files early.
Given this subject and my maximum all-in budget of [AMOUNT]:
[PASTE FACTS]
List conditions under which I should walk away before making an offer.
Include: ARV comp scarcity, rehab over [REHAB_CAP], hold over [HOLD_CAP] months,
HOA issues, title clouds, and if MAO would exceed [MAO_CAP]. Use bullet format.
If data is missing, say "cannot screen — need [MISSING ITEM]."Verify: Your own MAO math in the 70% rule calculator.
Screening rule of thumb: If the model returns more than two bucket-A walk-away flags before you have inspection data, pause. Either the lead is thin or you did not paste enough facts. Add records, re-run, then decide whether comps are worth the next hour.
Underwrite ARV, offer, and MAO: five prompts
These follow the comp rules in how to calculate ARV and MAO rules in the 70% rule guide.
5 — Sold-comp selection and ARV
When: You have pulled sold comps yourself.
You are a fix-and-flip underwriter. Subject after-rehab target: [BEDS/BATHS, SQ FT, FINISH LEVEL].
Sold comps I provide (do not add comps I did not paste):
[PASTE COMP 1–5: ADDRESS, SOLD DATE, PRICE, BEDS/BATHS, SQ FT, CONDITION]
Rules: use ONLY my sold comps. Drop any comp outside 0.5 mile unless I labeled it rural.
Drop comps more than 20% different in sq ft unless you adjust. Take MEDIAN adjusted value,
not average. Output: table of comps, adjustments, median ARV, and confidence (high/medium/low)
with reasons. Flag if fewer than 3 usable comps remain.Verify: MLS or agent CMA. Appraiser will use the same sold-data discipline.
6 — Adjust-and-median worksheet
When: Comps are close but not identical.
Using only these sold comps:
[PASTE COMPS]
Subject after rehab: [SUBJECT DETAILS]
For each comp, suggest dollar adjustments for: sq ft difference, bed/bath difference,
garage, basement finish, busy street vs interior lot. Show adjusted sold price per comp.
Take the median. List every adjustment as "investor estimate — appraiser may differ."
Do not use active or pending listings as comps.Verify: Your market’s actual adjustment dollars with a local agent.
7 — Classic MAO
ARV (from my analysis): [ARV]
Rehab (contractor estimate or my scope): [REHAB]
Calculate:
- MAO at 70% rule: (ARV × 0.70) − rehab
- Sell-side costs at 8% of ARV
- Buy-side costs at [BUY_COSTS] if I provide them
Show a table. State that the 30% leftover must cover profit, both closes, and hold.
Do not call MAO "guaranteed profit."Illustrative example (not a live deal):
| Line | Amount |
|---|---|
| ARV | $320,000 |
| 70% of ARV | $224,000 |
| Rehab | $70,000 |
| Classic MAO | $154,000 |
| Sell costs at 8% | $25,600 (from ARV) |
| Hold (5 mo, est.) | $12,000 — holding cost guide |
| Leftover 30% before hold | $70,000 — must cover sell costs + profit |
At $154,000 MAO, if the seller wants $175,000, the gap is not “negotiate harder.” Either ARV, rehab, or the 70% factor has to move — or you walk.
Worked example: five prompts on one file (illustrative)
Subject: 3/2 ranch, 1,380 sq ft, built 1982, Midwest suburb. You paste five sold comps from MLS: $305k, $318k, $322k, $328k, $335k (all updated resales, 0.3 mile). Median $322,000 ARV. Rehab bid $68,000. Hold 5 months at $2,400/mo = $12,000.
| Step | Result |
|---|---|
| Classic MAO | ($322,000 × 0.70) − $68,000 = $157,400 |
| 2026 MAO | $157,400 − $12,000 hold = $145,400 |
| Max loan at 75% ARV | $322,000 × 0.75 = $241,500 |
| All-in at MAO | $145,400 + $68,000 = $213,400 |
| LTC vs cap | 100% LTC on $213,400 = $213,400 — under $241,500 ARV cap ✓ |
If ARV drops 10% to $289,800, 75% ARV = $217,350 — still above all-in on this illustration. If rehab runs to $82,000 (+20%), all-in hits $227,400 and profit shrinks fast. That sensitivity is what prompt #9 is for.
8 — 2026 MAO minus holding costs
When: Interest-only carry is material.
ARV: [ARV]
Rehab: [REHAB]
Estimated hold: [MONTHLY_HOLD] per month × [MONTHS] months = [TOTAL_HOLD]
Loan rate for hold math: [RATE]% interest-only on average balance [AVG_BALANCE]
Calculate:
MAO = (ARV × 0.70) − rehab − total_hold
Compare classic MAO (without subtracting hold) vs 2026 MAO. Show both. Flag if spread
between them exceeds $10,000.Verify: Run the same file in the fix-and-flip calculator.
9 — Kill-this-deal / lender-fit screen
When: Before you submit to hard money.
Subject: purchase [PURCHASE], rehab [REHAB], ARV [ARV], hold [MONTHS] months.
Lender rules: max 75% ARV, max 100% LTC on qualified files, 6–12 month term.
Calculate max loan at 75% ARV and at 100% LTC (purchase + rehab). Fund the lower number.
List: (1) cash required at closing, (2) reasons the file fails if ARV drops 10%,
(3) reasons it fails if rehab runs 20% over, (4) exit if retail sale slips 60 days.
Do not approve the deal — list pass/fail against these rails only.Verify: Submit your file for a real leverage quote.
Rehab scope and holding costs: three prompts
Pair with how to estimate rehab costs and average rehab costs.
10 — Scope-to-budget from a walkthrough list
You are a rehab estimator. Finish level: [COSMETIC / MID / FULL GUT].
I will paste my walkthrough notes room by room.
Build a line-item scope table: item, quantity, unit cost RANGE (low–high), extended cost,
and tag each line BID REQUIRED or ESTIMATE. Include 15% contingency on subtotal.
Separate cosmetic from structural. Flag any line that requires permit or engineer as VERIFY.
Do not use national averages without labeling them "rough — bid required."Verify: Two contractor bids minimum on any file over $25,000 rehab.
11 — Unseen-systems and contingency
Subject: [YEAR BUILT, SQ FT, VISIBLE ISSUES]
Planned scope: [PASTE SCOPE SUMMARY]
List unseen-systems risk by category: roof, HVAC, plumbing (including polybutylene),
electrical panel, foundation, sewer lateral, mold/water. For each, probability
(low/med/high) and whether my scope already covers it. Recommend contingency %
for this age and market. Mark all dollar amounts as BID REQUIRED.Verify: Inspection with photos. Scope creep kills ARV timelines.
12 — Monthly carry stack
Average loan balance during project: [BALANCE]
Interest rate: [RATE]% interest-only
Annual property tax: [TAX]
Insurance quote: [INSURANCE]/month
Utilities: [UTILITIES]/month
HOA: [HOA]/month or $0
Project length: [MONTHS] months
Build a month-by-month hold table: interest, tax, insurance, utilities, HOA, total.
Sum total hold. Show daily burn (total ÷ days). Compare to my MAO leftover after
sell costs. Flag if hold exceeds 40% of expected gross spread.Verify: Insurer quote and tax bill. Use Jaken Finance Group’s published 8.99%–13.5% range for sensitivity if you do not have a term sheet yet.
Illustrative monthly hold (not a quote):
| Month | Avg balance | Interest at 11% | Tax | Ins | Utils | Total |
|---|---|---|---|---|---|---|
| 1 | $150,000 | $1,375 | $350 | $175 | $200 | $2,100 |
| 2 | $185,000 | $1,696 | $350 | $175 | $200 | $2,421 |
| 3 | $210,000 | $1,925 | $350 | $175 | $200 | $2,650 |
| 4 | $210,000 | $1,925 | $350 | $175 | $200 | $2,650 |
| 5 | $210,000 | $1,925 | $350 | $175 | $200 | $2,650 |
| Total | — | $8,846 | $1,750 | $875 | $1,000 | ~$12,471 |
That is why the 70% rule guide recommends subtracting hold explicitly in 2026 — the classic formula hides this inside the leftover 30%.
Rental and DSCR screens: four prompts
For hold exits, see hard money to DSCR refinance and the DSCR calculator.
13 — Long-term rent from your rent comps
You are a rental underwriter. Subject: [ADDRESS, BEDS/BATHS, SQ FT, PROPERTY TYPE].
Rent comps I provide (sold/leased — do not invent comps):
[PASTE 3–5 LEASE COMPS]
Rules: use ONLY my comps. Same product type and similar size. Output median market rent,
range, and confidence. If I provided fewer than 3 comps, say "insufficient — pull more comps."
Do not use Zestimate or a single listing ask as market rent.Verify: Form 1007, property manager rent survey, or lease in hand.
14 — Operating expense ratio
Market rent (verified): [RENT]/month
I will paste my expense assumptions: taxes, insurance, HOA, management %, maintenance %,
vacancy %, capex reserve %.
Build a monthly P&L and annual NOI. Show expense ratio (expenses ÷ gross rent).
Flag if any line is missing. Do not assume 50% rule unless I ask for a generic screen.Verify: Insurance quote and tax cert. STR expenses differ — see DSCR for Airbnb.
15 — DSCR worksheet
Monthly PITIA (or PITI if I provide): [PITIA]
Gross rent: [RENT]/month
Calculate DSCR = gross rent ÷ PITIA. Show at 5% and 10% vacancy haircuts.
State pass/fail at DSCR 1.0 and 1.25. Note: lenders use appraiser rent and their
payment quote — this is a screen only.Verify: DSCR calculator with your actual rate quote. Jaken Finance Group DSCR runs 5.75%–10.5% on qualified files.
Illustrative DSCR screen (verify rent and payment with lender):
| Input | Value |
|---|---|
| Gross rent | $2,050/mo |
| PITIA (quoted) | $1,640/mo |
| DSCR | 1.25 |
| At 10% vacancy haircut | $1,845 rent ÷ $1,640 = 1.12 |
A 1.12 might work on some DSCR programs; others want 1.0 minimum on appraiser rent, not your model. Prompt #15 is a screen — the DSCR loan appraisal and 1007 is the file.
16 — STR vs LTR regulation checklist (not a revenue promise)
Subject: [ADDRESS, CITY, COUNTY, HOA IF ANY]
Strategy options: long-term lease vs short-term rental
Build a CHECKLIST ONLY (no revenue forecast): zoning/STR allowed?, HOA rental/STR caps?,
local registration/permit, occupancy tax, insurance type (landlord vs STR rider),
neighbor complaint risk, seasonality flag. Tag each item VERIFY WITH [CITY/HOA/INSURER].
Do not estimate nightly rates.Verify: City STR ordinance, HOA CC&Rs, insurer STR endorsement.
Wholesale screens: three prompts
For wholetail timing, see wholetail financing.
17 — Assignment-fee room after buyer MAO
Contract price with seller: [CONTRACT]
End buyer profile: fix-and-flip, 70% rule
End buyer ARV: [ARV]
End buyer rehab: [REHAB]
End buyer hold: [HOLD]
Calculate end buyer MAO = (ARV × 0.70) − rehab − hold.
Assignment fee room = end buyer MAO − contract price − buyer closing cushion [CUSHION].
If fee room < [MIN_FEE], say NO DEAL. Show math in a table. Do not assume a fixed
15% ROI rule — use the buyer's MAO.Verify: Buyer’s actual criteria and proof of funds.
Illustrative wholesale math:
| Line | Amount |
|---|---|
| Contract with seller | $120,000 |
| Buyer ARV | $280,000 |
| Buyer rehab | $55,000 |
| Buyer hold | $10,000 |
| Buyer MAO | ($280,000 × 0.70) − $55,000 − $10,000 = $131,000 |
| Cushion for buyer close | $3,000 |
| Assignment room | $131,000 − $120,000 − $3,000 = $8,000 |
If your minimum fee is $10,000, prompt #17 should return NO DEAL unless ARV, rehab, or contract price moves.
18 — Unsellable-deal checklist
Subject: [PASTE FACTS — TITLE, ACCESS, OCCUPANCY, FLOOD, HOA, PRICE]
List reasons a cash flip buyer or landlord would refuse this deal despite a low price.
Cover: title clouds, probate without authority, severe fire/flood, no legal access,
condo litigation, HOA rental ban, environmental, over-improved for area, comp desert.
Tag each as VERIFY WITH title/attorney/HOA. Do not provide skip-trace steps.Verify: Title preliminary and buyer feedback.
19 — Wholesale one-page buyer packet draft
Turn my verified facts into a one-page buyer email outline (no hype):
Subject: [ADDRESS]
As-is price: [PRICE]
ARV (source: my comps): [ARV]
Rehab range: [REHAB_LOW]–[REHAB_HIGH] BID REQUIRED
MAO for 70% buyer: [MAO]
Known red flags: [PASTE]
Photos/inspection: [AVAILABLE YES/NO]
Assignment terms: [FEE] fee, [EARNEST] earnest, closing [DAYS] days
Use bullet format. Label every number I must still verify. No owner phone numbers.Verify: Every number before you blast buyers.
How should you finance the deal the model just sketched?
AI does not approve your loan. It helps you assemble the package faster. When the worksheet looks real, map it to Jaken Finance Group programs:
| Program | Rate range | Leverage | Term | Close |
|---|---|---|---|---|
| Fix and flip / hard money | 8.99%–13.5% | Up to 100% LTC on qualified files; up to 75% ARV | 6–12 months | 7–10 business days |
| DSCR rental | 5.75%–10.5% | Up to 85% purchase, 80% cash-out on qualified files | 30-year fixed or ARM | 14 business days |
See 100% LTC program details and 100% financing overview. Not sure which product fits? Start at what kind of loan do you need.
Plan A vs Plan B exit (prompt the model on both)
Retail sale is Plan A on most flips. Plan B is what keeps you out of a fire sale: DSCR refinance if the house rents, wholetail to another investor, or a modeled price reduction with extended hold. Add this block to prompt #20:
Also sketch Plan B if retail DOM exceeds 90 days: hold cost for 3 extra months,
price reduction scenarios at -3% and -5% ARV, and whether DSCR refi is plausible
given rent comps I provide (do not invent rent).Lenders underwrite the exit you can prove today. A Plan B paragraph in ChatGPT does not replace a lease or buyer list — it forces you to ask the question before you are stuck on month seven.
Documents the lender package prompt should never skip
- Sold comp sheet (addresses, dates, adjustments)
- Contractor bid or detailed scope with contingency
- Insurance quote (vacant or builders risk)
- Entity docs if borrowing in an LLC
- Exit summary: list price strategy or DSCR path
- Photos or inspection if available
AI can draft the outline. You attach the evidence.
20 — Lender package prompt
You are a hard-money loan packager. I will paste my verified deal summary:
Purchase: [PURCHASE]
Rehab: [REHAB]
ARV: [ARV] (source: [COMP COUNT] sold comps)
Hold timeline: [MONTHS] months
Exit: [RETAIL SALE / DSCR REFI]
Borrower experience: [DEALS COMPLETED]
Produce a lender submission outline: executive summary (5 sentences), comp summary table,
scope summary, hold table, exit strategy, top 5 risks, and documents checklist
(appraisal, insurance, entity docs, contractor bid). Flag gaps. Do not invent borrower
credit or loan approval.Next step: Submit your flip file or call (833) 264-7776. Jaken Finance Group lends in all 50 states.
Verification protocol: eight checks before you offer
Run this on every AI-assisted deal:
- ARV — 3+ sold comps you pulled; median matches agent or appraiser bracket
- Rehab — contractor bid or line-item scope; contingency included
- Hold — month-by-month carry at your actual or quoted rate
- MAO — (ARV × 0.70) − rehab − hold; compared to 70% rule
- Rent — lease comps or 1007, not a model guess
- Title — preliminary or attorney review; no AI lien clearance
- Insurance — vacant or builders risk quote in hand
- Leverage — file fits 75% ARV and LTC rails; run fix-and-flip calculator
If any check fails, the prompt output is a draft — not a deal.
Daily workflow: where each prompt fits
| Stage | Prompts | Time saved |
|---|---|---|
| Lead in inbox | #1–#4 screen | 15–20 min vs ad-hoc notes |
| Offer decision | #5–#9 underwrite | 30–45 min vs blank spreadsheet |
| Scope meeting | #10–#12 rehab/hold | Aligns GC bid to MAO |
| BRRRR or rental | #13–#16 DSCR | Catches weak rent before close |
| Wholesale blast | #17–#19 buyer packet | Fewer “what’s the ARV?” replies |
| Lender submit | #20 package | Cleaner file, fewer back-and-forth emails |
Run verification after each block — not only at the end. Catching a bad ARV at prompt #5 saves every downstream hour.
Common mistakes new investors make with AI
- Asking for comps instead of pasting comps — the model will invent addresses
- Using list price as ARV in the prompt — you baked in the error
- Skipping hold in MAO — double-digit IO rates make this fatal in 2026
- Trusting “motivated seller” language — require HYPOTHESIS tags (prompt #3)
- Submitting AI output as the appraisal — lenders need your comp sheet and their appraiser
- One prompt for the whole deal — chain the templates; each step has different constraints
Want a downloadable prompt cheat sheet? Request our ChatGPT for Real Estate guide (email capture). For live deal help, submit your project or use what kind of loan do you need to pick the right program.
Sources
- Topic prompt only (not republished): PropStream — 60+ Real Estate AI Prompts, May 26, 2026. We did not copy its one-line question list or product claims.
- AI adoption and impact: NAR 2025 Technology Survey press release, September 18, 2025 (survey fielded July 2025; n=1,241). AI content use 46%; 46% no noticeable business impact; ChatGPT 58% among AI tool users cited in release.
- ARV, MAO, rehab, hold methodology: Jaken Finance Group guides linked above (ARV, 70% rule, rehab, holding costs).
- Loan terms: Jaken Finance Group published program parameters (fix-and-flip 8.99%–13.5%, up to 100% LTC and 75% ARV on qualified files; DSCR 5.75%–10.5%).