Scrub and cluster the list before you spend a stamp
Rung 2 of 10 · Comparison matrix
1 · Learn the move · Comparison matrix
A raw pulled list is full of noise: duplicates, undeliverable addresses, owners who died, corporate entities mixed with individuals, and the same owner scattered across ten rows. Mailing it blind wastes most of the spend. The move is a comparison matrix over the list YOU paste: one row per record, columns that tag the record, out-of-state, likely-deceased, entity-vs-individual, duplicate, and a cluster key that groups every parcel one owner holds so you can make a single portfolio offer instead of ten postcards. The discipline that keeps this honest is that every tag is a HYPOTHESIS, not a fact. 'Likely deceased' means the pasted data shows a pattern worth checking against the records, never a conclusion, and it becomes a verification step, not an action. The AI works only the rows you paste, never enriches from memory or invents a status, and it makes no inference about an owner as a person, only about the data quality and the parcel pattern in front of it.
Clean and cluster this owner list. Flag duplicates, out-of-state, entities, and group parcels by owner. [paste the list rows]
2 · Your turn. You write the prompt
You pulled 400 rows and you're about to mail. You paste a sample: several rows share the same owner name and mailing address (a portfolio), two rows have owner names with 'LLC' and 'TRUST', one row's owner shares a name with a mailing address flagged 'c/o Estate of', a few have out-of-state mailing addresses, and there are obvious duplicates. You want the list scrubbed and clustered so you mail once per owner and skip the junk, without the AI inventing an owner status or enriching data you didn't paste. Write that prompt.
Remember: the AI sees only your prompt, not this page. If the situation isn't in your prompt, it doesn't exist.
Optional. These shape the output when you run your prompt below, not your score.