← BACK TO THE BLOG

The whole fabrication workflow, mapped.

I’ve run a stone and fabrication shop. The machines were never the bottleneck. The saw doesn’t lose jobs. The CNC doesn’t forget to follow up. What eats the week is everything between the machines: the enquiry that sits, the quote that goes unchased, the change that never makes the invoice.

So here’s the whole workflow, mapped. Three fixes, in order, with what each looks like before and after.

Stage 1: Enquiry to job sheet (automate)

Before: a photo and a vague email sit in the inbox for two days while someone chases dimensions, material and a deadline. The customer, who messaged three shops, already has a number from one of them.

After: every enquiry gets an intake link back within minutes. Three questions, a spot for photos, under a minute on a phone. An AI step drafts the answers into a job sheet in your format and flags what’s missing. Ready to price the same day. I’ve written this one up in full in its own post if you want the detail.

Stage 2: Quote follow-up (automate)

Before: quotes get chased from memory between jobs, if they get chased at all. Nobody knows which ones are going stale until someone reads the whiteboard by accident.

After: follow-ups send themselves on a schedule you set, in your words, until there’s an answer. The moment the customer replies, a human takes over. Remembering is not a skill. It’s a schedule.

Stage 3: Variations and rework (systemise)

Before: the customer rings mid-job, asks for the shelf 100 deep instead of 60, someone says “yeah, no worries”, and the change never touches the job sheet or the invoice. The shop eats it, or argues about it after handover.

After: every change gets logged as a variation before the work happens. What changed, which materials, how much time, whether the deadline moves. The invoice matches the job, and the awkward conversation happens up front, while it’s still small. AI can draft the entry from a phone note:

PROMPTVARIATION ENTRY
Turn this note into a variation entry: what changed, materials affected, extra time, and whether the deadline moves. Leave the price impact blank for me to set. If the note doesn't say who approved the change, flag that at the top.

But be honest with yourself here. This one is mostly a habit with a form attached. The discipline fixes it. The AI just makes the discipline cheap.

What gets left alone

The cutting. The fitting. The measure on site. The phone call where you talk a customer through stone versus laminate. Anything where your judgement or your hands are the product stays exactly as it is. In most shops, that’s more than half the workflow, and it should be.

The machines were never the bottleneck.

The order matters

Start with intake, because it wins jobs you’re currently losing to slow replies. Then follow-ups, because they compound the intake win. Then variations, because they protect the margin on everything the first two bring in.

One workflow at a time. Straighten it, hand it to the team, move to the next.

No revolution required. Just fewer leaks between the machines.

Want this mapped for your shop? The audit does exactly this on your real jobs, in a week.
BOOK A FIT CALL
Shane Black
WRITTEN BY Shane Black.

Twenty-odd years running hands-on businesses, including a stone and fabrication shop. Now builds practical AI systems for owners on the Gold Coast, and writes up what works as he goes.

Keep reading.

All posts →