AI-Powered Operating System for Commercial Roofing

Bringing AI to a $92B Industry (All Segments)
That's Still Running on Spreadsheets

ABEPIM is a 5-layer operating system that transforms how commercial roofing contractors estimate, propose, execute, and manage projects — replacing days of manual takeoff work with AI-driven precision.

The Problem We're Solving

Today's Reality

  • A full commercial estimate takes 4–12 hours of manual blueprint analysis; the takeoff alone consumes days on large bid sets
  • Spreadsheet-driven takeoffs prone to human error
  • No real-time visibility into project costs or progress
  • Reactive maintenance — fix it after it leaks
  • Disconnected tools for estimating, proposing, and execution
  • The estimators who do this work by hand are retiring — and there's no pipeline replacing them

The ABEPIM Solution

  • AI-powered blueprint analysis delivered a measured 20-sheet extraction in ~28 minutes on a real 120-page bid set (expert review follows)
  • Multimodal vision models extract exact quantities from drawings
  • Live project dashboards with real-time cost and timeline tracking
  • Scheduled maintenance programs prevent failures before they happen
  • One unified platform — estimating through closeout, fully integrated

One Platform, Five Layers

ABEPIM isn't just an estimating tool. It's a complete operating system for commercial roofing — each layer builds on the last, from the first blueprint scan to autonomous project management. Nine roadmap modules (including payment rails; all roadmap) extend the platform into code compliance, insurance integration, roof asset management, permit automation, aerial verification, sustainability tracking, and a building owner portal that creates a two-sided industry platform.

1

AI Estimation Operational Today

Multimodal vision models analyze blueprints, extract material quantities, and generate takeoffs — measured at ~28 minutes of AI extraction for 20 sheets on a real 120-page bid set, with source-linked results available for estimator review for expert review. Already running on live commercial roofing bids.

2

Proposals & Pipeline Operational Today

Auto-generate branded, professional proposals from takeoff data. Track every bid from submission to close with a unified pipeline dashboard.

3

Project Execution Year 1–2

Live project tracking, crew coordination, progress reporting, and change order management — all connected to the original estimate and proposal.

4

Financial Operations Year 2–3

Invoicing, margin tracking, revenue recognition, and financial reporting tied directly to project data — closing the loop between estimation and actuals.

5

Autonomous AI Agent Year 3+

An AI agent that monitors projects, flags risks, generates reports, and proactively manages workflows — operating as a virtual project manager across the entire platform.

Built, Not Theoretical

ABEPIM's core estimation engine is already operational on live commercial roofing projects — not a concept, not a mockup. A working prototype processing real bids today.

$300K
~28min
Measured AI extraction, real 120-pp bid set (review follows)
$300B+
Multi-trade expansion market (all MEP, TAM estimate)
$499/mo
Entry pricing, scales to $2,999/mo

"As a Florida-licensed roofing contractor with 10 years in the field, I've seen every tool promise to 'revolutionize' estimating. ABEPIM is the first one that actually delivers. It pulls measurements, calculates materials, and generates a complete proposal in minutes — what used to take my team days. ABEPIM isn't just software — it's the difference between us bidding on a handful of jobs a month and bidding on twenty-plus. I've been doing this long enough to know when something is a gimmick and when something is a game-changer. This is a game-changer."

— Cory Williams, Owner, America's Best Roofing and Construction (FL License CCC1332167)

Why It's Defensible: Harness Engineering

A better prompt improves one answer. A better harness improves every run. The model is commoditized — every competitor rents the same frontier. ABEPIM's defensibility is the six-layer execution system built around the model, which turns probabilistic output into verified, deliverable takeoffs.

6
Execution layers: task contract · context compiler · permissioned tool gateway · durable state · evidence gates · trace & recovery
$0/page
Measured marginal cost of the default plan reader — prose sheets read from the drawing's own text layer with no model call
21/21
Facts recovered on a blind spec-sheet instrument, text-first path, 0.03 s (second instrument: 20/21)
~10×
Faster than manual development, measured by OpenAI on an agent-first internal build (published Feb 2026)

OpenAI's own account of that build is the clearest statement of where the value sits: zero manually written lines of code, roughly one million lines and ~1,500 merged pull requests within five months — and the part they called interesting was not that the agent could write code. It was what the engineers had to build around it before the code became useful. That is the harness. A competitor can copy a prompt in an afternoon; they cannot copy six layers of enforcement without operating the business the software is for.

We measure the harness by one number: reviewable work produced per unit of human attention.

Meet the Team

J

Jason van Joosten

Founder & CEO · USAF Veteran

U.S. Air Force veteran with 10+ years roofing industry experience (residential + commercial). Former Regional Manager at SeekNow overseeing $20M+ in annual revenue across FL, GA, and AL. Building ABEPIM from deep firsthand knowledge of commercial roofing workflow bottlenecks.

C

Cory Williams

Operations Partner & Pilot Customer

Licensed FL contractor (CCC1332167). 50/50 partner across all operating entities. Brings years of residential and commercial roofing field experience and an active contractor network across Florida.

W

Wesley Schlemmer

Technical Lead

B.S. Chemical & Biological Engineering, Tufts University. Founder of Crossthebridge.io (2 years, AI agent development across commercial roofing, commercial appraisal, and film production). Architect of the ABEPIM prototype — already operational on live bids. Personal builds include Batou.ai and Nomus.studio.

N

Nashville Automation Co.

Technical Build Partner

Full-stack AI engineering firm. Built production AI for construction (Pristine Acres). 8+ active clients. Founders Rob Castro (COO) and Matthias Debernardini (CEO/CTO). LOI signed for engineering partnership upon funding.

D

Dan Cain

Senior Advisor

Former VP & GM at SAS (6+ years leading AI/analytics business operations). Currently Chief Cybersecurity & Analytics Officer at Reef CyberSecurity. Advises on cybersecurity strategy, governed AI adoption, and enterprise sales — de-risking ABEPIM's path to SOC 2 certification and Fortune 500 contractor adoption.

Let's Build Something

Interested in partnering, investing, or learning more about ABEPIM?

jasonv@abccfl.com
Investor Deck One-Page Summary