End-to-end teardown of how Imperium Building Solutions produces a facade bid, mapped against where Facade.ai takes over. Reverse-engineered from the 80 West Broadway package (RFP → scope doc → leveling form → drawings).
The market gap, in Arianna's own words: facade firms use takeoff software to measure quantities — but you still manually identify which areas to measure, and nothing facade-specific carries those quantities into estimating (price generation + a knowledge database). Our wedge: automate the area-identification takeoff software can't do, then extend past quantities into the estimating realm — the part no tool currently touches.
GC emails an RFP + a Dropbox link. Simone opens an all-disciplines design package and pulls out only the facade-relevant sheets & specs by hand.
Ingest the entire package, classify every sheet/spec, and surface the facade subset automatically. Pure pattern-work — no judgment needed.
Read the GC's Exterior Finish Schedule (every material code → source, color, location) and the SOW/pricing template to establish what's to be furnished.
Extract the full TAG# → material → location table and parse the SOW line items into a structured scope list.
Simone prints elevations, hand-colors every surface by material, counts units, measures m²/lm, and applies assembly rules. Takeoff software measures once you highlight — but you still pick what to highlight.
Auto-identify which regions are which material (the step takeoff tools can't do), extract counts + dimensions, and apply the codified assembly rules.
The same quantities live in 3 formats — architect (imperial) → Imperium SOW (metric) → GC bid form (imperial/bid categories). Simone converts & cross-checks by hand, every revision.
Deterministic conversion + category mapping + self-validation. Because the numbers must agree across documents, the system can verify itself with no human check.
Email the correct material subset to each supplier (window assembly mfr, glass, terracotta, shipper), then wait days while each redoes the takeoff on their portion.
Auto-package the right material set per supplier from Stage 2–3 output, send, and track responses. Gated on a stable supplier roster (unknown).
Meet with suppliers' engineers; flag missing info; optimize design for production → raise RFIs to the GC. (Takeoff tools auto-generate RFIs — Imperium doesn't use them.)
Pre-empt missing info via a per-component checklist (the recurring "assumptions") and draft RFIs. The checklist piece could land early — human meetings stay human.
Supplier prices arrive; Imperium layers in overhead, PM, customs & duties, contingency, engineering stamps, Italy travel → the final numbers.
Mostly hands-off — this is Imperium's judgment/margin logic. Exception: a historical $/SF lookup for the early ballparks GCs ask for (needs a price database — unknown).
Type every quantity into the GC's leveling template and re-format into whatever per-element breakdown each client requires.
Populate the leveling form from Stage 3 output and generate the per-client breakdown. The output schema already exists (below) — high value, low difficulty.
A new drawing set lands every few weeks for 6–12+ months. Simone manually hunts the deltas between old and new sets and re-prices. "Un bagno di sangue."
Diff drawing versions, flag exactly which quantities changed, and re-run Stages 2–3 only on the deltas. Strong recurring-value play once v1 exists.
| Scope category | Unit | Qty (80 W Bway) | Furnish + Install |
|---|---|---|---|
| Punched windows (tilt & turn, incl. glass, receptors, straps) | SF | 8,761 | $1,267,703 |
| Painted aluminum window surround | LF | 3,491 | $314,627 |
| Storefront system | SF | 912 | $237,967 |
| Terracotta panels @ storefront | SF | — | $285,200 |
| Cantilever + mech-screen metal panels | SF | 3,269 / 3,660 / 655 / 566 | $1,360,762 |
| Engineering, calcs, submittals, design-assist | LS | — | $291,162 |
| Subtotal — Furnish & Install | $3,984,619 | ||
The output format is fixed and learnable — we don't invent it, we populate it. A structured Alternates block sits below (2-side vs 4-side receptors +$50K/+$128K, terracotta-vs-metal surround −$105K, NY PE stamp +$29K, mockup allowances, bird glass excluded). Caveat: read from cached values, not live formulas — get the formula-level file before building against this schema.