A timed talk track, the four numbers to memorise, what to say when they push back, and what not to do. When the slot is short, the deliverable is not the document — it's the talk track plus a live link.
"We run LendingPad and Floify. Nacho fuses them into one platform we own, feeding our own pricing engine. Not a gap-filler — the system. And because we own it, we can sell it."
"It costs an IMB $12,209 to make a loan and we earn $727 on it. Production expenses went up last quarter. At that margin, every touch we remove is real money — and per-seat licensing taxes exactly the growth we're chasing."
"Blue Sage is good. They ship embedded AI today — document extraction, mismatch detection, condition automation. So 'we'll have AI' is table stakes, not our edge. Our edge is the stuff nobody builds: the worksheets our operators run in Excel today, and owning both halves of the borrower-to-back-office seam." (Saying a competitor is good buys you credibility for everything after it.)
"This isn't a wish list. It's 236 numbered features, each with acceptance criteria you can demo, built from Sharon's operational notes and checked against the code that already exists. And it flags ten things Sharon asked for that the current blueprint dropped — including the brokered channel, which is one loan in eight."
"Two decisions block real progress. One: who owns disclosure delivery — three documents currently say three different things, and whoever owns it owns the three-day deadline. Two: the credential list. DU, LPA, MERS, FHA Connection — CLEAR sources those, not the developers, and they gate any real end-to-end test. Then hand them your phone and let them tap through the screens."
| They say | You say |
|---|---|
| "Isn't this just what Sharon already sent?" | "Hers is the operational truth and it's good — I built on it. Mine adds the shipping layer: compliance as controls instead of a category list, the QC program, the architecture, the migration, and the seam to the POS. Plus ten things from her own notes that didn't survive into the blueprint." |
| "Why not just buy Blue Sage?" | "That's a real option and the document argues both sides honestly. Buying transfers risk. Building buys upside — the worksheets and the seam aren't purchasable, and per-seat pricing taxes our growth. But we should ask Blue Sage one question first: do they allow a third-party POS? That answer changes the math." |
| "Can we go live end of August?" | "As a demo, on a controlled staff test file in a sandbox — yes, and that's worth doing. As a real loan — no, and I'd rather say so now. A supervised pilot is months, gated by credentials we source. Production is 12–24 months. The danger is mistaking 'it ran through the demo' for 'we're almost done.'" |
| "How is this different from the AI everyone's announcing?" | "In an LOS, AI suggests — it doesn't decide. And since August 6th, Fannie requires a written AI governance framework and can demand to know what the model saw and who reviewed it. We built to that from the start, which is also what makes it sellable to other lenders." |
| "Who's actually going to build this?" | "The VERONTECH team, against this spec. That's the point of the format — every feature says where it lives, how it looks, and how you know it's done. There's a 236-item backlog they can score against the existing code." |
| "What's it going to cost?" | "The document doesn't invent a number — and I'd be suspicious of one that did. What it does give you is the phasing, the risk register, and the credential list, so the cost conversation happens against real scope." |
The single most likely way this project loses credibility. Say "demo" every time you mean demo.
Parts of it are still operator testimony, and it's labelled that way on purpose. Show the tagged VERIFIED rows only.
It's the least exciting slide and the most likely to blow the timeline. Owners and dates, in the room.
Part 8 needs counsel review before build. The document says so; don't accidentally promise more.