How Scope Architect works
The terminology, the math, and the controls — explained. Read top-to-bottom once, then come back when a number doesn’t look right.
Overview
Scope Architect turns raw discovery notes into a granular, billable project scope. It’s a single-operator tool: you input what the client said, the AI structures it into a tree of work, and you calibrate, edit, and export it as a proposal.
The output exists in two parallel forms — an Internal Blueprint (every task, every assumption, every hour) and a Client Proposal (polished, outcome-focused, hides anything you flagged internal-only).
Workflow
- Anchors— pick from preset module categories (Branding, Web, SaaS, CRM, AI). These “lock” foundational tasks the AI cannot remove.
- Calibration — set complexity, risk buffer, discount, and timeline pressure (crunch adds a rush premium).
- Discovery — paste your call notes or bullet-point dump. Name the project. Hit generate.
- Blueprint workspace — the AI streams modules in. You edit, toggle visibility, flag risks, recalibrate live.
- Export — switch to Client view, hit Export Proposal, pick a theme, save as PDF.
Modules → Features → Tasks
Every project is a three-level tree. Understanding the level you’re editing matters because it changes what “granularity” means.
Pricing is summed bottom-up: tasks have hours → features have hours → modules have hours → the project total is the sum of all modules with global multipliers applied.
Effort Levels
Every task gets an effort level — a label that maps to a number of raw hours. Your hour constants live in Agency DNA — defaults shown below.
| Level | Default hours | Use when… |
|---|---|---|
| XS | 2 hrs | Trivial — config tweak, copy update, single-file change |
| S | 4 hrs | Half-day — small component, simple integration, polish pass |
| M | 8 hrs | One day — typical feature, standard CRUD, moderate API integration |
| L | 16 hrs | Two days — multi-step flow, complex state, third-party deeply integrated |
| XL | 32 hrs | A week — major subsystem, novel pattern, heavy infrastructure |
Risk Levels
Each task has a risk level. Tasks marked high risk get a flat +20% bufferapplied to their hours. Low and medium don’t add anything individually — they’re informational, used by you to decide where to add the project-wide risk buffer.
The Shield icon next to each task in the blueprint toggles between medium and high — the fastest way to flag a risk during review.
Tiers — Foundation, Studio, Signature
Every task is tagged with a tierthat says “at which proposal level this task starts being included.” This drives the multi-tier proposal — three good/better/best documents from one underlying scope.
In the blueprint, every task shows a small circular tier marker (F / S / ★). The dropdown next to it lets you change the tier inline.
Preview tier
The sidebar has a 4-segment tier preview switcher (All / Foundation / Studio / Signature). When set to anything other than “All”, the blueprint dims out tasks excluded from that tier and the price/hours numbers reflect only included tasks.
In the proposal
The proposal page has two render modes:
- Single — render the proposal at one tier. The toolbar tier picker lets you flip between Foundation, Studio, and Signature without leaving the page.
- Compare— side-by-side cards for whichever tiers you select. A “Show” toggle in the toolbar lets you include any combination of Foundation, Studio, and Signature. Studio is highlighted as “Recommended” when shown alongside others. Clients pick the middle option ~60% of the time when offered three — the single biggest conversion lever in agency proposals.
Calibration Engine
Calibration controls the four global multipliers that turn raw task hours into billable hours and price. Live in the right sidebar, expandable under “Calibration.”
How the Math Works
The full pipeline, from a single task to the final invoice:
rawHours = sizeConstant × (riskLevel === "high" ? 1.2 : 1.0)Example: a Medium task at high risk = 8 × 1.2 = 9.6 hoursprojectRawHours = Σ taskRawHoursSum of every task's raw hours across all modules.billableHours = rawHours × complexity × (1 + riskBuffer/100) × timelinePressureThis is what shows in the “Total Effort” stat. Timeline pressure is 1.0× for relaxed/standard, 1.15× for crunch.totalPrice = billableHours × hourlyRate × (1 + discountPct/100)Discount is negative or zero, so it only ever reduces. Rate uses override if set, else Agency DNA.timelineWeeks = Math.ceil(billableHours / 40)Assumes one full-time equivalent. Adjust mentally for parallelism.Worked example
A project with 80 raw hours, Complexity 3 (1.25×), 15% risk buffer, Standard timeline (1.0×), $150/hr, no discount:
Agency DNA
Your global defaults — applied to every project. Access from the top nav (gear icon) or the full /settings page. Changes save to Supabase so they survive across devices and deploys.
Anchors & Presets
Presets are pre-built “Golden Standard” module templates. When you anchor one, the AI is forbidden from removing the tasks inside — it can only expand and refine. Anchored items show a small pin icon in the blueprint.
The pin icon appears next to anchored modules, features, and tasks in the Internal view. Clients never see this — it’s purely an internal cue about what was “forced in” vs. what the AI generated freely.
Internal vs Client View
The view toggle in the sidebar swaps the entire workspace between two perspectives:
Every task has an Eye icon that toggles its visibleToClient flag. Use it for internal-only work — research spikes, refactor passes, tech debt — that you don’t want on the client’s document.
Manual Editing
Every text field in Internal view is editable. Click any title, description, or note → it becomes an input. Save with Enter (Cmd+Enter for multiline), cancel with Esc, or click away to save.
Export & Proposal
The Export Proposal button takes you to /scope/proposal — a dedicated client-facing document. Three themes available:
The Print / Save PDF button uses your browser’s native print dialog. The page is configured for US Letter with smart break rules:
- Cover page is always full-page
- Modules flow naturally across pages
- Assumptions + Exclusions stay paired
- Investment Summary + Sign-off forced to its own final page
- Headings never get orphaned at the bottom of a page
Saving & Projects
Scopes are saved to Supabase — not just your browser. This means they survive device switches, deploys, and cleared caches.
/settings. Changes there don’t auto-save — use the “Save Agency DNA” button or the Save button in the Agency DNA panel drawer.Keyboard Shortcuts
| Enter | Save inline edit (single-line) |
| Cmd + Enter | Save inline edit (multiline / textareas) |
| Esc | Cancel inline edit, restore previous value |
| Tab | Move between inputs / toggles when editing |