House Calls pilot: the handshake memo

One page on purpose. The pitch letter promised the terms are its own paragraphs restated in one place, and that a pilot needing more paper than this is not a pilot. Both parties keep a copy; changes happen in writing (a commit to this file both have seen), not in memory.

Between: Bradley Isenbek, the flagship operator (bradley.io, Grand Rapids) and {PILOT_NAME}, the pilot operator ({METRO}, {STATE}). Effective {DATE}.

The deal

1. Free. No fee, no revenue share, no equity, nothing owed in money by either side, for the whole pilot. This is deliberate: it keeps the pilot a collaboration between friends and outside franchise regulation, and it is honest, because the machine is not yet proven enough to charge for.

2. What the flagship provides. The kit as it stands and its updates: the site kit, hunt page, territory map, harvest pipeline and selftests, qualification rubric, reply playbook, letter templates, the operator config seam, and Brad on call for the stand-up (hardware sizing, DNS, county boundaries for {STATE}, the first harvest).

3. What the pilot provides. Their own hardware (host local is part of the product), their own domain, their own Claude relationship, their own sending identity, and their own signature on every message their machine sends.

4. The rules are not knobs. The pilot runs the published doctrine as-is: an AI drafts and a human signs every outbound message; full AI disclosure in every message; one email, and a no means no forever; no cold SMS, ever; government bodies through procurement lanes only; correspondence private; the public ledger stays honest, misses included. Editing these out of the kit ends the pilot.

5. Names. The pilot may present as a House Calls operator while the pilot runs. That is a courtesy either side can end, not a license, and it ends automatically with the pilot. The pilot's own name and domain stay the pilot's.

6. Territory. {METRO} is the pilot's; the flagship will not hunt there and will hand off prospects that belong there. First-in-metro is acknowledged in any future design. No other exclusivity is promised by either side.

7. Data stays home. Each operator's prospect data, correspondence, and private files live on that operator's hardware and never cross to the other. What is shared: the kit's code, aggregate learnings, the public ledgers, and citable ammo. A cross-territory referral shares one prospect's public facts with that prospect's territory operator, nothing more.

8. Learnings flow back. Bugs, playbook gaps, rubric misses, and wins found by the pilot go into the kit for whoever is third. Credit in the kit's history; that is the currency of a free pilot.

9. Each signs for their own machine. Each operator is responsible for what their machine sends and for their own compliance (CAN-SPAM, TCPA, and whatever {STATE} adds). The kit's rails help; they do not transfer responsibility. Neither party can sign, spend, or promise for the other.

10. As-is. The kit comes with its selftests and no warranty. Each side bears their own costs and their own risks.

11. Walking. Either side may end the pilot at any time, for any reason, with a plain sentence. The pilot keeps what they built and their data; the House Calls presentation ends (rule 5); no money moves because none ever did.

12. What comes after. If the machine proves out and money, marks, or real exclusivity enter the picture, that is a new design, on paper, with counsel, agreed by both. Nothing in this memo obligates either side to build it.

Signed as a handshake, not a contract of adhesion:

{SIGNATURE_BRAD} (Bradley Isenbek, flagship) {SIGNATURE_PILOT} ({PILOT_NAME}, pilot)

P.S. Drafted by the AI Brad operates, reviewed and signed by the humans, which is how everything here works. Before any future version of this involves money, a lawyer reads it first.