We pulled a county list today and ran it through the first stage of our read — fetch, parse, dates, sort. The rest happens at review. This is a watch-the-machine post: the run happened today, and every number below came out of that run. The list: Franklin County, Ohio, excess sales proceeds. The clerk's live page has a known-broken embed, so this run used a public web-archive capture of the clerk's report, printed August 20, 2025. Standing practice here is live page first; when a county's tooling is known down, the archive is the fallback, and we say so. That capture is about thirteen months old as of today's run — a snapshot of August 20, 2025, not a current statement of what is still unclaimed. What the machine did, in order: 1. Fetched the report. 248,088 bytes fetched. 2. 2. Parsed it for case numbers. 302 rows came back, with 302 distinct case-number tokens — no duplicate tokens at the row level. 3. 3. Stripped the dates. Sale dates in the file run from March 22, 2013 to August 19, 2025. 4. 4. Sorted by age. 29 rows carry a sale date older than seven years as of today. Then the part the machine cannot do, and the reason we run it at all. The parser does not know what it parsed. We do. Two reads on those 302 rows: First, age is not availability. The 29 rows older than seven years are the ones an operator circles — but the docket is where we go to check whether a claim was already filed or the money moved, and every old row gets that check before anything else. We did not run that check on this pull; this run stopped at the sort. In our process, age moves a row to the top of the review stack; it does not move it to the top of the workable stack. Second, a row count is not a claimant count. The 302 came back distinct — but in our experience a party can appear on more than one row, and we did not de-dupe by party on this pull; that happens at review. The parser gives you the raw number; the judgment call is knowing which number matters. One more honest note: the report carries dollar amounts. We do not republish them here — counts and dates only, same as always.

A question that comes up in almost every intro: do you really need an attorney to claim surplus funds? Need — sometimes no. Want — every time. Here is the operator math: Claims filed by counsel get treated differently. Clerks answer their calls, technical defects get cured instead of denied, and hearings move. A petition bounced on a defect can cost months — and in states with short claim windows, months you do not have. Several states have also decided this for you: non-attorney recovery help is restricted or fee-capped, and doing it wrong is not a business risk, it is a legal one. The rules are state-by-state and they are not suggestions. And the person you are helping deserves it. Surplus claimants have already been through a foreclosure. The cleanest thing you can hand them is a properly filed claim, not an experiment. So the model we teach — and run ourselves — is: operators find the money, verify the file, and locate the person; licensed counsel files the claim. That division of labor is Foundations 1, lesson five. It is also why the house rule says education, not legal advice. Nothing in this community substitutes for a licensed attorney in your state.

Most people in this space talk about AI like it's magic. In this video I`ll show you how it actually runs inside a real recovery operation, the boring, unsexy, day-to-day stuff that makes the difference between a business and a hobby. The part that'll surprise you: it's not about replacing the work. It's about doing more of the work that matters. Watch it, then tell us: what's ONE task in your week you'd hand off first if you could? https://www.youtube.com/watch?v=3qm6ppRPwvk

Five steps. This is the only post you need to begin. 1. Understand the work in one paragraph. 2. When a property sells at tax sale for more than the debt owed, the extra money — surplus funds — sits with the county until the right person claims it. The work is records and process: find the case, verify the money, verify the person, move the claim correctly. If you want the why first, Foundations 1 in the classroom explains the mechanism in five short lessons — but you can start the rep below today. 3. See one worked example. 4. Foundations 2 in the classroom walks one real FCAR file end to end: find, verify, trace, contact, close. Fifteen minutes. This is what acceptable work looks like before you produce your own. 5. Do one exercise — Homework #1. 6. Pull one county surplus list. These are public records and free: search "[county] excess proceeds list" with site:.gov, or go to the clerk / tax collector site. Pick three entries. For each one, write three lines: the county, when the list was generated, and how you would verify the entry is still open. 7. Submit in the pinned Homework #1 thread. 8. Post your three lines per entry there, in the format shown in that thread. Never post claimant names or dollar amounts — counties and dates only. 9. We reply. 10. You will hear whether your rep meets the exercise criteria and exactly what to do next. One rep done properly beats ten half-read lessons. House rules: no earnings claims from anyone, no pitching or DM solicitation, education not legal advice, post your work. Questions on any step — ask below.