An Official Member
COXIT builds custom AI/ML software that help your team bid more projects, catch revision changes faster, and stop digging through 500-page spec books by hand.
We review how your team handles drawings, documents, specs, and revisions, identify where time is lost, and define the first automation worth building.
We build AI trained on your actual drawings and documents – not off-the-shelf tools. Private, secure, integrated with the systems your team already uses. You keep the IP.
We maintain the system and add capabilities your team requests. The accuracy is monitored continuously – the system gets better the longer you use it.
→ Faster Quote Turnaround
Quote more projects with the same team.
Combination of AI and Machine Learning drops manual quote review time from 2-6 hours to about 10 minutes.
See the case ↗
→ Catch Scope Changes
AI compares revisions and surfaces what changed – no more flipping between PDFs by hand.
→ Reduce Paperwork
AI finds the right section in about 30 seconds with exact page references.
→ Catch Errors Early
Compare them side by side, linked by elevation reference – mark issues, track status, generate audit reports.
→ No More Retyping
No manual transcription, no copy-paste errors.
→ Get AI-Ready
Most shops run on a mix of legacy estimating tools, Access databases, and internal systems nobody wants to touch.
We map what you have, migrate what matters, and rebuild what’s holding you back.
We built two AI systems on their actual production drawings and specs. Stevens owns and operates both today.
Estimators used to spend 2–6 hours per project reviewing drawings by hand — counting cabinets, cross-referencing notes, transcribing into the quoting system.
The system we built reads the blueprints, detects cabinetry, cross-references MasterFormat specs, and outputs a structured estimate ready for their quoting workflow. Live in production at Stevens since 2025.
Finding the right product series in a spec book used to take 30+ minutes of scrolling.
Now their team types a question and gets an answer with page and paragraph citations in under 30 seconds.
Handles large files. Live in production at Stevens since late 2025.
We start by looking at what your team actually works with – drawings, specs, quoting outputs, ERP exports. Not every shop needs AI on everything.
We figure out which problem to solve first, and whether AI is the right way to solve it.
Send us drawings, specs, and quoting outputs. We review them and tell you where AI applies and where it doesn't.
Working model on your actual data. Accuracy metrics, a demo on your drawings, and a roadmap for production.
Full system - web app, integrations with your quoting tools, training for your team. Ongoing support retainer available.
meet the expert
Book a consultation and discover how you can improve your organization.
LinkedIn
// Interview By Volodymyr Hresko
Hear directly from Curtis Garrard at Stevens Industries about how AI completely changed the way they handle estimates. Learn about the real problems they solved, how the team adapted to new technology, and the results they’re seeing today. Get the full story — and some practical advice—straight from someone who’s been through it.
Frequently Asked Questions
Send us 2–3 of your recent project PDFs. Within 3–5 days we send back an annotated drawing showing what our AI detects, a structured spreadsheet of items found, and an honest read on whether AI is the right fit for your workflow. No commitment, no sales pressure. If it’s not a fit, we’ll tell you.
The honest answer: not until we test it on yours. Off-the-shelf tools fail on production drawings because they’re trained on synthetic data or rely on text search for AWI codes.
We trained custom systems with real drawings and hit 91% accuracy. Your drawings are different — different title blocks, different conventions, different cabinetry mix.
That’s why every engagement starts with a Free Technical Review on your actual files before anyone signs anything.
Your drawings and specs stay in isolated environments — not used to train models for other clients, not shared, not pooled. Deployments run on AWS or GCP under your account if you prefer, or under ours with full audit logs. We sign NDAs before any file moves. For shops with stricter requirements, on-prem deployment is possible.
You own what we build. The system, the trained model on your data, the code – all yours, in your repository, from day one. We retain methodology and knowledge, not your system.
Free Technical Review: free. Discovery + Proof of Concept: $15K–$35K, 4–8 weeks. Production build: $80K–$150K, 3–6 months. Ongoing support: $2K–$5K/month. Most shops start with the PoC to validate accuracy on their drawings before committing to a full build.
Yes. We export structured data (CSV, JSON, or direct API) that drops into Cabinet Vision, Microvellum, Innergy, or custom ERPs. Output formats are defined during discovery so the spreadsheet matches your quoting workflow line-for-line. No retyping, no copy-paste.
First annotated output from Free Technical Review: 3–5 days. PoC running on your data with accuracy metrics: 4–8 weeks. Production system in daily use: 3–6 months. Stevens went from manual review to AI-assisted estimation in under 6 months.
Tell us a bit about your project or challenge, and we'll get back to you shortly.
Contact our co-founder
We've received your details and will be in touch within 24h. In the meantime, feel free to explore our work.
01// What do you need help with?
02// What’s your name?
03// What’s your email address?
04// Provide project details:
