Deal intake comes first: a complete inventory record for every closed deal

· How to use this site

  1. Understand done
  2. Explore done
  3. Convergenow
  4. Commit
Next stepSeed the inventory with the 48 deals already run through intake
StatusGroup leaning: intake first, before the other jobs
Waiting onWaiting on Reid and Joe
Last movedOct 19
ProjectExperiment 1: four ways to build the co-worker, code-baseline

Decisions 4

D9 Deal intake end to end is one flow round 1 of the experiment tests (closed deal folder in, complete inventory record out, with proof for every fact and questions instead of guesses); other flows may be added, because Intake is the first gate; the later steps all trust what intake writes. Answers already exist for the closed deals.
Agreed by Joe · accepted · from Oct 19 · Joe, Reid, Keegan
D8 Seed the inventory now with the data we have and build the processes on top of it; fill the data gaps later as each process gets solved, because Reid: we have no real data to build on yet, and chasing perfect first stalls everything; once ingest and the emails work, it's a data problem, not an operations problem. Keegan: agreed; keep pushing on the app and get the data as the problems get solved.
Agreed by Reid, Keegan, Joe · accepted · from Oct 19 · Joe, Reid, Keegan
D7 Round 1 of the experiment tests deal intake end to end: closed deal folder in, complete inventory record out, with proof for every fact and questions instead of guesses, because Intake is the first gate; the later steps all trust what intake writes. Answers already exist for the closed deals.
Agreed by Joe · replaced by D9 · from Oct 19 · Joe, Reid, Keegan
D5 Deal intake (closed deal into a complete inventory record) is the first job to solve, before the other jobs, because Keegan: the inventory is the brain of the operation; every later step depends on it. Reid: the gate into our system is after close.
Agreed by Keegan, Reid · accepted · from Oct 19 · Joe, Reid, Keegan

No questions, to-dos, requirements, drafts or files yet.

History · 9

Mon, Oct 19, 2026 · Joe, Reid, Keegan

  • Added to the board
  • Keegan: the inventory is the brain of the whole operation. Nothing else matters until every closed deal lands in the inventory as a complete set of facts. Reid agreed: the gate into our system is after a deal closes, so intake is the first thing to solve, and we are close. Joe: the core tools are in place; we need to pull more facts per record (business rules) and hand it to the AI. Reid noted paperwork also arrives before close (title packets), so scanning is not only after transfer.
  • Moved from Understand to Converge
  • Now feeds code-baseline
  • Reid: the inventory does not have to be 100% backfilled before we start; we can take in new deals and do their transfers now. Keegan: going forward, every new deal should be as close to 100% complete at intake as possible, so nothing needs fixing later. The one extra thing intake must do is the basis allocation for accounting (how the price paid splits across each property). Reid: that is already part of intake. Keegan does the basis allocation by hand today for every deal closed this year; the accounting team should do it, but the data isn't good enough yet.
  • Reid: the intake process has run on 48 deals so far. The first experiment matters because it unlocks taking in more documents and seeding the inventory. The goal for the AI workers is to lower what needs a person for everything, not hand Kelsey, Tana or Kathryn a new pile of work. Joe: the AI co-workers get their own computers; we talk to them and they work on a schedule like any employee.
  • Joe's shortlist of AI co-worker versions so far: OpenHands, Claude Code, NemoClaw (NVIDIA's locked-down OpenClaw), Letta (built on the MemGPT research from Berkeley, with long-term memory), and Hermes Agent (works with many models and comes with memory and learning built in). What he is pushing for in every version: it stops and asks when it is confused instead of guessing. The scoring is nearly done; what's left is settling the versions.
  • On checking the answers for the closed deals: Reid's earlier test asked whether those answers are actually right, since they were built from HubSpot and the Excel for 40 to 50 deals. Keegan cautioned against trusting the Excel and HubSpot, because people typed those in by hand; trust what's on the recorded title documents (the PDFs of record) instead. Reid agreed. Joe and Reid will settle the AI versions together.
  • Reid: going forward, every document that comes in gets read and structured from the recorded papers themselves. But to build the intake tool now, we use the existing closed-deal answers: they may not be 100% right, but they're right in direction, and that's enough to build a tool that knows what it's looking for.