type: reference status: active created: 2026-06-06 owner: Reid + Arden anchor: reference/core/the-nuance-gap.md related:

• decisions/2026-05-17-practice-builders-membership-model.md
• decisions/2026-06-06-leveraged-pipeline-and-bd-role.md
• prospects/martin-eisenstein/profile.md ---

The Custom Build Playbook

How to scope, sell, and deliver custom AI build work — fused from the strategy chair (Reid) and the builder's chair (Arden), anchored by The Nuance Gap.

Use before and during any engagement where you're building something specific to one owner's business (Deploy Sprint, AOS build, bespoke agent). the-nuance-gap.md is the principle; this is the operating manual.

The anchor (one paragraph)

The commodity is the agent; the value is the owner's nuance — the judgment they make on the fly and have never encoded. You can't buy it off the shelf; you surface it by building and testing, and that iterative loop is the engagement. Everything below operationalizes that.

Part 1 — SELL (position + scope)

• Sell the externalization, not the agent. They think they're buying "an AI agent" (commodity, sounds DIY-able). They're actually buying their tacit process turned into an owned, runnable asset. Lead with that and it stops competing with "I'll figure it out myself."
• Name the Nuance Gap out loud: "What's in your head isn't a system until it's encoded and tested. That's the work — and you can't shortcut it." Justifies the engagement, the price, and the timeline in one move.
• Shape the offer to the owner (discriminators):
• Participation surface area — how much sustained showing-up/implementation does it demand? High-demand vehicles fail with low-participation owners.
• Done-FOR vs. done-WITH vs. requires-their-work — handing a draft to a non-implementer fails (it dies undeployed). Built-and-running-inside-the-engagement, with their implementer, lands.
• Recurring vs. project — read their buying history (some commit to ongoing monthly and balk at lump sums; some the reverse).
• The real test under all of it: will this get to "running," given who they are?
• Set the iterative expectation up front: "We'll have a v1 fast, then test it on your real cases and encode what only shows up when it runs. Plan about a month." Kills the later "why isn't it done / why does it cost this" friction.
• Don't justify your price or validate their payment. Peer to peer. State scope and investment plainly; no coaching language.

Part 2 — DELIVER (the build mechanics)

• The build is days; the engagement is iteration + their review cadence. Don't sell "6 weeks of building" — sell "running in weeks, hardened on your real cases."
• The loop: build v1 → run on real data → they react ("no, you'd never do X, because Y") → encode Y → re-run. 4–6 cycles. Their reaction is the extraction — the rule only surfaces when they see the wrong output.
• Builder builds async, owner judges sync. Heads-down build between sessions (fast); the live block is react-and-encode only. Their async is the enemy, not yours.
• The live weekly session is the compression mechanism, not the delay. Sacred calendar time to present, get feedback live, kill snags in real time — no inbox lag (where deferrers vanish). Walk in with a runnable thing that just produced real output on real data.
• Ship a visible v1 early (session 1–2) — results-in-advance; they see it work, which de-risks doubt and helps the sale.
• Build WITH their implementer (whoever will maintain it) and end at "working & tested" — so implementation happens inside the engagement, not handed off to die.
• Resolve the integration fork early (the Pro Connect lesson): design to read a clean export/feed; don't depend on operating someone else's UI (fragile, possible ToS issue). Settle it in session 1.
• Compliance/infra is day-one — licensed pros + sensitive data → Teams/Enterprise before real data touches it.
• Maturity-model anchor: can't build Level 2/3 (AI assists / runs) without Level 1 (it's documented). The build forces the missing Level 1 into existence.

Part 3 — MANAGE THE HUMAN (change management)

• Surfaced gaps are the map, not failures. When the agent stalls, that's where their process was never specified. Frame it as "here's how we make it real," never "the tool failed."
• Sequence the gaps: let them discover some by watching it run (creates buy-in); flag-and-decide others (saves a cycle). Discovery teaches; too much discovery is slow.
• Get them on board with constraints ("we can't do it the way you pictured, because X"). This is where the advisor, not the builder, earns the seat — trust/relationship makes "your process has a hole here" land instead of offend.
• Forecast by their participation + realization speed, not the build. Responsive owner → compresses to ~2–3 weeks. Deferrer → needs the weekly container and a longer realization runway. (See the Martin profile for a worked forecast.)

Appendix — The content vein (raw angles → route to Sloane)

This framework is dense with positioning + social content. Seeds, not finished posts — content production is Sloane's lane; these are raw angles to develop, each mapping to the offers (Sprint/AOS) and the products:

• "You can't buy an AI agent off the shelf and have it work your way."
• "The commodity is the agent. The value is your nuance."
• "Your process isn't a system — it's you. Take yourself out and the gaps appear."
• "Assistant vs. delegated team member: most people have never been forced to write down the judgment they make on the fly."
• "Why a custom AI build takes a month, not a day — and why that's the point."
• "The build is a forcing function: it fails exactly where your process was never real."
• "Async is where the work dies. The live session is where it compresses."
• "You don't buy an agent. You buy your business turned into something that runs without you."
• Positioning line (ties to the brand): "I don't teach you about AI. I turn what's in your head into a system that runs." — capability deployed, not curriculum learned.

→ Next step: Sloane mines this appendix + the-nuance-gap.md into LinkedIn posts, positioning copy, and offer-page language.