# Tom Brannick, Owner / Head Estimator at Brannick Precision — read of ManufactureAI, June 10 2026

> 19 years in CNC machining, started as an operator, bought the shop in 2018. 18 people. Mostly aerospace and defense small-batch work out of Dayton, Ohio. Still write probably 60% of the quotes myself.

## How I got here

Google. I searched "job shop quoting software small shop" last Tuesday after we lost a repeat customer to a shop in Indiana that got their bid in same day. We quoted the same job in three days. I've been meaning to look at this stuff for a year. I finally had a reason that stung. Clicked through maybe six sites before I hit this one.

## What I clicked first

The headline: "Slow quotes cost precision shops real jobs every quarter." Yeah. That's the thing. Not a vague productivity pitch. That specific word "quarter" hit me because I actually do think about it quarterly. Every three months I look at lost jobs and a chunk of them are timestamped faster competitor.

I kept reading past the hero, which I almost never do.

## Where I paused

The bit about the 22-person machine shop. "ManufactureAI was built for a 22-person machine shop where the owner is also the head estimator and the scheduler is pulling double duty on the floor." I stopped on that. That is my exact situation. My scheduler is also running our Haas VF-4 three days a week. The specificity there felt earned, not made up.

The Ray Kowalski quote also made me stop. "Cut our quote time from two days to three hours." Two days to three hours is a number I can sanity-check against my own shop. Two days is about right for a complex job. Three hours sounds aggressive but not impossible if the system is pulling from real historical data.

## What I distrusted

The header stats. "93% on-time delivery improvement." That number is absurd unless I know what it's measuring. Improvement from what baseline? Whose shops? How many? That stat looks like it was designed to be impressive rather than to mean something. "4x faster quote turnaround" is the same flavor of number. I've seen enough vendor decks to know these figures come from either one cherry-picked customer or a methodology that doesn't survive scrutiny.

Also: "Unlock the dossier · $5." What is that? There's a $5 dossier teased in the hero and then it disappears. I have no idea what I'd be buying. That's a weird enough choice that I spent thirty seconds confused about whether this was a normal SaaS product or some kind of information product. Not a great first impression.

And: no actual pricing on the page. The nav says "Pricing" but what I got from the homepage is nothing concrete. "Book a Free Demo" is the only call to action. That's a pattern I've learned to read as "it's expensive and they want to qualify you before they say the number."

The testimonials are fine but they're also the three testimonials every one of these pages has: the owner who is winning more jobs, the production manager who saved his Monday mornings, and the front office person who stopped answering the phone. I'd believe one of them more if there was a company website I could click through to.

## What would convince me

One thing: show me the actual AI estimating step on a real part. Not a mockup, not a screenshot. A two-minute screen recording where someone uploads a PDF drawing of a turned part and I can see what the system actually generates. Does it read GD&T callouts? Does it know the difference between a 5-axis and a 3-axis job? Does it get confused when the drawing is hand-marked? That tells me more than all five of those stats combined.

Second thing: a real quote from a real shop that I can compare to what the shop would have done manually. Even an anonymized one. If you have a customer who will let you show side-by-side what their old Excel quote looked like versus what ManufactureAI output, that closes the question for me.

## What I'd ask in an email reply

1. The AI estimating piece pulls from "historical job data and current material prices" -- what does that mean in practice for a shop that's been using spreadsheets? My historical data is in 11 years of Excel files with inconsistent formatting. What does the actual import process look like and what do you need from me to make the cost model usable on day one?

2. How does the system handle jobs with long lead times on raw stock? A lot of my aerospace work quotes in January for material I won't buy until March. The page talks about flagging margin risk from price changes, but how does the lock/alert feature actually work when the customer insists on a firm fixed price up front?

3. What's the actual monthly cost for a shop my size? Not a range, not "let's talk." I'm not booking a demo to find out you start at $800 a month.

## Verdict: curious-enough-to-reply

The 22-person shop line and the quote turnaround problem are described well enough that I believe the people who built this have actually talked to shops like mine. If I get a straight answer on pricing I might actually keep reading.

---
*Memo by skeptic persona, generated 2026-06-10. Studio breaks own self-grading loop.*