The AI estimating workspace for builders & quantity surveyors

Accurate estimates,
kept accurate by AI.

Material is a multimodal AI workspace that turns drawings, CAD files, spreadsheets, documents, emails, call transcripts, images, video, site photos, chat, and voice notes into one live estimate. The AI reads every signal, reasons over the full project, and keeps the commercial picture current as the project evolves.

Multimodal AI ingestionSource-backed AI reasoningLayered confidence (L1–L4)Margin clarity
material/Coastal house · Hawthorne Rd · WIP● synced 2m ago
Sources
L1.dwg
L2.dwg
scope-v3.pdf
est-baseline.xlsx
arch_re_window.eml
site-east-wall.jpg
walkthrough.m4a
Email sync
14 new from architect
3 new from client
Estimate · v14
$1,284,500 ± 6.2%
GFA · AVG
312 m² · $4,116/m²
L1·CAD
L2·DOCS
L3·TEAM
L4·TRADES
Item
Qty
Unit
Rate
Total
Carpentry
$284,200
Subfloor framingCAD
312
$148
$46,176
Window header — east wallAI·v14
2
ea
$1,840
$3,680
Roof trussesCAD
22
ea
$640
$14,080
Excavation
$58,900
Difficult-access surchargeAI·v14
1
ls
$8,200
$8,200
Bulk excavationCAD
184
$92
$16,928
Joinery
$112,400
Kitchen carcassesMAN
12
lm
$2,200
$26,400
Refine
You · voice note
“Site has rocky access on the east side — only walked-in skid steer can fit, and there’s heritage fence we can’t damage.”
Material
Detected L3·Team + L4·Trades evidence: difficult-access + heritage constraint.

Updated 3 rows; estimate moved +$11,880 (0.9%). Confidence on excavation: 68 → 81.
Sources cited
walkthrough.m4a · 0:42–1:08
site-east-wall.jpg
L1.dwg · sheet A-201
DWG·DXF·IFC(roadmap)·PDF specs·XLSX estimates·Architect emails·Site photos·Voice notes·Call transcripts·Trade comments·Site video(roadmap)·Chat threads·DWG·DXF·IFC(roadmap)·PDF specs·XLSX estimates·Architect emails·Site photos·Voice notes·Call transcripts·Trade comments·Site video(roadmap)·Chat threads·
Any source · one estimate

Project information rarely arrives in one clean package.

It arrives across drawings, spreadsheets, specs, email threads, site photos, voice notes, chat messages, and trade inputs. Material’s AI absorbs all of it — reading drawings, parsing documents, transcribing voice, interpreting photos — so every signal improves the same estimate, instead of starting another disconnected thread.

L1.dwgCAD
site-plan.dxfCAD
scope-v3.pdfDOC
spec.docxDOC
est-2024.xlsxXLS
RE: revisionsEML
client briefEML
east-wall.jpgIMG
rebar-detail.jpgIMG
walkthrough.mp4VID
voice-note.m4aAUD
trade-call.wavAUD
thread #joineryCHAT
supervisor msgCHAT
subbie quoteTRADE
site obs.TRADE
ONE LIVING ESTIMATE
$1,307,580
v14·± 6.2%·● 4 layers
A single project · five hours

Watch the AI absorb a day’s reality.

A coastal house, three weeks into pre-con. Drawings, an architect email, a supervisor’s photo, a builder’s voice memo, a trade thread — each lands in the workspace, and the AI updates the estimate with source-backed reasoning. No reconciliation by hand.

Builder · 9:42 AM
L1.dwg, L2.dwg, est-baseline.xlsx
DWG drawings + spreadsheet estimate uploaded
L1·CAD+$1,284,500
Baseline generated · 6.2% range
Living estimate · v9L1·CAD
Subfloor framing312 m²CAD$46,176
Roof trusses22 eaCAD$14,080
Bulk excavation184 m³CAD$16,928
L1
60%
L2
0%
L3
0%
L4
0%
auto-playing
How it works

Four steps. Nothing more.

Material is built around a workflow that mirrors how estimates actually get built — and re-built — through the life of a project.

01

Bring in project information

Upload drawings, CAD files, spreadsheets, specs, photos, audio, and documents. Connect Gmail to sync architect and client threads. Paste in trade comments, chat, and site notes.

02

AI generates a baseline estimate fast

Material’s AI extracts CAD-derived quantities, parses your existing spreadsheet logic, and produces a grouped estimate by trade — with descriptions, units, rates, totals, feasibility range, and confidence metadata. Minutes, not days.

03

The AI keeps refining as new information arrives

An architect emails a scope change. A supervisor uploads a site photo. A builder records a voice memo. A trade weighs in. Every signal is assessed by the AI against accumulated project context — not just the latest message.

04

See confidence, reasoning, and margin impact clearly

Inspect any line item: AI-written summary, detailed reasoning, references, and last-updated metadata. Confidence is layered (L1 CAD · L2 Docs · L3 Team · L4 Trades). Margin is tracked as scope evolves.

What’s in the workspace

The capabilities that make a living estimate possible.

Material is more than a calculator and more than an assistant. It’s a continuously maintained commercial model of your project, with the evidence right there.

Core CAD & plan analysis

Quantities directly from drawings.

Parse .dwg and .dxf. Extract areas, dimensions, entities, and drawing-derived scope. AI-generated line items linked back to the sheet they came from.

Trust Source-backed reasoning

Every change traces back to evidence.

Open any line item to see why it changed and what it changed from — with references to drawings, specs, emails, photos, voice notes, and chat.

L1·CADL1.dwg · sheet A-201
L2·EMLRE: revisions · 10:18 AM
L3·IMGsite-east-wall.jpg
L3·AUDwalkthrough.m4a · 0:42
Differentiator Layered confidence

Confidence isn’t a single number.

L1 CAD, L2 Docs, L3 Team, L4 Trades — each layer reflects what evidence supports the estimate. Maturity is visible at a glance.

L1·CAD
86
L2·DOCS
72
L3·TEAM
58
L4·TRADES
34
Memory Context & anti-recency bias

Remembers the full project, not just the latest message.

Site conditions, design complexity, builder notes, open questions, assumptions — all accumulated, all checked against the next signal that lands.

heritage fence
rocky access
arched openings
polished concrete
heritage overlay
east-wall tie-in
double-height stair
tight street access
Conversational Chat-driven refinement

Ask, attach, adjust.

Talk to the estimate. Drop in a photo. Record a voice note. Material updates the rows that actually changed and tells you what moved and why.

“Add the heritage flashing detail from the east-wall photo.”
Updated 2 rows · est. moved +$2,400
Integrations Gmail · Xero

Plug into the threads where decisions actually live.

Sync project email through Google Workspace OAuth. Push approved estimates into Xero as quotes and invoices.

Gmail · architect@studio.cosynced 4m ago
Gmail · client@buildco.ausynced 4m ago
Xero · Q-2046draft created
Margin clarity

As scope changes, margin shouldn’t.

Material connects technical scope and source evidence directly to estimate confidence, feasibility range, total cost, and margin risk — so you can see what changed, why it changed, how confident the estimate is, and what it means before it costs you.

Margin posture · todayHeld
14.2%
▲ 0.3%since baseline · v6 → v14
10:18Architect email+$3,680−0.1%
11:04Site photo+$2,400−0.1%
13:36Voice note · access+$8,200−0.3%
15:12Trade · joinery+$6,800−0.2%
16:40Rate re-bid · steel−$5,200+1.0%
Security & trust

Sensitive drawings. Handled with care.

Material is built for serious commercial work. Source visibility is the trust mechanism — explainable AI, project-scoped access, and signed media URLs are the substrate.

01

Project-scoped access

Every drawing, document, email, photo, and audio file is bound to the project it belongs to. Access is checked at every read.

02

Encrypted in transit, signed at rest

TLS for every connection. Signed URLs for media access. OAuth-based Google integrations — your inbox stays your inbox.

03

Explainable, source-backed AI

Every estimate change references the evidence behind it. Confidence is layered, not hidden. Uncertainty is shown, not smoothed over.

See how Material keeps your estimate accurate from first drawings to final scope.

45-minute walkthrough · bring a real project · we’ll estimate it together