Where AI Actually Fits
Five places AI is starting to change steel cutting workflows — and the honest gaps that remain.
By Justin — ProtoForge Technologies
Most of the AI conversation in manufacturing is happening at the vendor level — software companies adding features, consultants selling transformation. What's actually useful to a quoting desk or a programming department is a different question. These five areas aren't theory. They're drawn from the real friction points I've seen across fifteen years of CNC programming and workflow management. Some of this is available today. Some is close. And I've tried to be straight about where the gaps still are.
Assembly Drawing Interpretation
The Problem You Felt
That 34-part assembly PDF sitting in the inbox. Two hours of cross-referencing, line by line. Your salesperson doing detailed engineering work just to build a quote.
What's Emerging
AI tools that read engineering PDFs, extract parts and processes, and flag confidence levels. Instead of checking every line, your salesperson focuses on the uncertain items.
What's Available Now
Paperless Parts for complete AI quoting workflows. Mavlon for metal fabricator-specific extraction. Tempus Tools (US$100/month) for geometry extraction from vector PDFs. Claude or Gemini on business plans as a first-pass review tool.
The Honest Gap
Nobody has fully solved spatial reasoning — cross-referencing BOM items to specific parts in assembly views by following balloon callouts. It's coming, but it's not there yet.
A Bevel Knowledge Base
The Problem You Felt
Your programmer getting pulled off a 200-part nest to answer a bevelling question. Two people's workflow interrupted for one query.
What's Emerging
AI configured as an internal knowledge base, loaded with your shop's actual bevel rules — thickness limits, feasibility guidelines, material allowances.
What's Available Now
Claude or similar AI tools configured with your custom knowledge. Your salesperson queries it like asking a colleague. The programmer only gets involved for genuinely edge-case questions.
The Honest Gap
Requires someone to document the rules in the first place. The knowledge often lives in one person's head — getting it written down is the real work.
Work Order Review Before Handoff
The Problem You Felt
Missing information blowing up at the programming stage on a tight turnaround. Something the salesperson didn't catch, something the customer didn't provide.
What's Emerging
AI that reviews the actual work order and prompts with specific questions rather than a static checklist that gets ignored.
What's Available Now
Custom AI evaluators that check sheet sizes against part dimensions, flag unusual material grades, and catch common oversights before they reach the programmer.
The Honest Gap
Needs integration with your quoting system to access job data. Off-the-shelf solutions don't exist yet — this is custom build territory.
KPI Tracking and Nonconformance Analysis
The Problem You Felt
Monthly meetings where everyone looks at a spreadsheet of recuts and goes off gut feel. No pattern recognition, no actionable specifics.
What's Emerging
AI-powered analysis that finds patterns in your recut data, NCRs, and production issues. Specifics instead of vibes.
What's Available Now
Feed your data into AI tools for pattern recognition. The data usually exists somewhere — in your bespoke software, spreadsheets, or recut logs. AI pulls it together into something actionable.
The Honest Gap
Data quality. If your logging is inconsistent or incomplete, AI can't find patterns in data that doesn't exist. The first step is often fixing the data capture.
Procedure Generation
The Problem You Felt
You need SOPs. You know you need SOPs. But writing them is painful, and the result is usually a 40-page document nobody reads.
What's Emerging
AI that walks you through creating a procedure by asking the right questions, then produces something short, direct, and followable.
What's Available Now
Any business-plan AI tool can do this today. The key is the prompting approach — guided conversation rather than 'write me an SOP'. Results in practical guides that match how your shop actually operates.
The Honest Gap
The procedures are only as good as the knowledge input. AI can structure and format, but it needs someone who knows the work to provide the substance.