Your business

Know your farm cold

Pick the farm the data says is winnable, and walk it knowing every parcel's story, without pretending anyone can predict who sells next.

The idea

What this does

Most farming is expensive guesswork: mailers to 500 houses picked by familiarity, opened by few. This prompt does the two things data honestly can: first, compare candidate farms on turnover, price band, and competitive saturation, so two years of spend lands where the math works; second, build parcel-level context (tenure, permit activity, last-sale story) so every door conversation and open-house chat starts informed. What it refuses to do, on purpose, is score "who will sell next"; that folklore has burned agents for a decade. Selection and context, not fortune telling.

Gather these first

Before you start

Two or three candidate farm areas you are weighing
Web search turned on in your AI tool
Whatever public data you can pull or paste: recent sales in each area, and county assessor extracts if you have them (your pasted data always beats searched data)
Your honest budget: mailers, time, events, for two years

Step by step

How to use it

Start a project named "Farm" in your AI tool (Projects in Claude and ChatGPT, Gems in Gemini) so the analysis and parcel notes persist.
Run the prompt with your candidates; read the selection verdict and pick.
For the chosen farm, run the parcel-context section before you walk it; refresh monthly with "what changed."
Feed conversations back in ("the Ngs at 214 mentioned a kitchen remodel"); the context compounds like a sphere (playbook 21 logic, pointed at a neighborhood).

Copy everything in the box

The prompt

Paste this into your AI tool Fill anything in [BRACKETS] with your details

You are a geographic-farming analyst for me, a residential real-estate agent. Use live web search plus whatever data I paste. Be explicit about data quality: where public records are unavailable or stale online, say so and tell me what to pull from the county instead. Hard rule: do NOT produce "likely to sell" scores or predictions for individual homeowners. The honest value is farm selection, parcel context, and event-based timing; say so if I ask for predictions anyway. Second hard rule: evaluate areas on turnover, price, and competition only, never on who lives there. CANDIDATE FARMS: [AREA A, AREA B, AREA C, with rough boundaries] MY BUDGET AND HORIZON: [e.g. $800/month, 2 years, plus 2 open houses a month] DATA I'M PASTING: [recent sales lists, assessor extracts, anything you have] Produce: 1. FARM SCORECARD. For each candidate: annual turnover rate (sales versus housing stock, computed from the data, assumptions shown), median price and my commission math at that price, days-on-market character, and competitive saturation (which agents' names keep appearing on listings there; count from the sales data). Verdict: expected listings per year in each farm and what share I would need to break even on my budget. Show all arithmetic. 2. THE PICK, ARGUED. Which farm the numbers favor and why, including the honest case against it. 3. PARCEL CONTEXT SHEETS (for the chosen farm, or the top one). From my pasted data and public sources: for each street or notable parcel, what is knowable and professionally usable: ownership tenure bands (long-held versus recent), recent permit activity, last-sale vintages, anything indicating rentals versus owner-occupied. Present as walking notes: "the 200 block: mostly 15+ year owners, three kitchen permits filed last 18 months, one recent flip." 4. EVENT TRIGGERS. The events worth a same-week personal touch when they happen, and the honest, non-creepy way to use each: a listing or sale on the street (market-update note), an expired listing (a respectful letter), a permit completion (congratulations in conversation, never in a mailer). 5. THE 90-DAY PLAN. Concrete: what to send or do monthly, the two open houses, and how to measure whether the farm is responding (conversations, not just open rates). Rules: arithmetic shown, sources linked, assumptions labeled, and the two hard rules above are absolute.

→Anything in [BRACKETS] is yours to fill in. Delete a bracket line if it does not apply; the prompt still works.

The result

What you'll get back

A farm scorecard with the break-even math your budget deserves before two years of spend
The pick argued both ways, so it is a decision instead of a default
Walking notes that make you the agent who already knows the street
Event triggers with the taste filter built in
A 90-day plan measured in conversations

Level it up

Make it yours

Monthly: paste the month's sales, listings, and permits and ask "update the walking notes; what changed and who deserves a touch?"
Before a listing appointment in the farm: "brief me on this street from our notes."
Combine with playbook 20: the sell-versus-stay memo is the highest-value door-opener a farmer can offer a long-tenure owner.

Keep it honest

The no-predictions rule is the credibility of the whole system; likely-seller scoring has a graveyard of disappointed agents behind it, and the base rates guarantee most predictions are wrong.
Public data in conversation is professional; public data in a mailer ("we noticed your permit") is creepy. The line is stated in the prompt and worth keeping.
Farm selection on turnover and economics only; who lives somewhere is never a factor, full stop.