TRIVIALOS FOR MANUFACTURING
How many RFQs did you turn down last month? Quote them all.
A quoting workspace that reads the package, drafts the routine, and flags what makes the job hard. The agent inside is a helper: it asks when it isn’t sure, and it learns how your estimators judge.
30 minutes. Send a real RFQ beforehand and we run it live on the call.
One screen, the whole job.
From RFQ in to quote out in one place: the package, the part, the precedent, and the draft. The agent fills in the routine; your estimator makes the call.
PACKAGE
PUMP_HOUSING_CAST · REV D
EXTRACTED
LINE ITEM
Pump housing, EN-GJS-400-15 · Rev D
DRAFT PRICES
Lead time 9 weeks
COST BREAKDOWN · PER PIECE AT 2,400
AGENT
Read 3 documents. Material EN-GJS-400-15, 2,400 pcs a year, delivery in 12 weeks.
The spec changed in the last email. Rev C or rev D?
This screen isn’t fixed. It’s built on trivialOS, so when your team wants different fields, another view, or an extra step in the flow, you describe the change and it ships.
Sound familiar?
An RFQ arrives and there are two people in the building who can say whether you should quote it. Both are busy. One retires in four years and the judgment goes with him.
Requests sit in an inbox until they go stale. Nobody decided to decline them. You just ran out of week.
You’ve made something like this before. Finding out what it cost means asking around, or opening folders named by year.
Quoting software wants one tidy STEP file. Your customers send a scanned drawing, three PDFs, and an email that changes the spec in the last paragraph.
It starts by not knowing your shop. That’s the point.
Every quoting tool on the market ships with someone else’s assumptions baked in, usually a machine shop’s. Then it meets your foundry drawing and starts guessing.
Ours starts empty on purpose. On the first RFQ it asks a lot: the minimum wall for this alloy, the draft you assume, which of your plants runs this size. Your estimator answers once, and it doesn’t ask again. Over a few weeks it builds a picture of how your shop decides. That picture is yours. It doesn’t go into a shared model and it doesn’t leave your building.
It asks about everything
Reads the package, extracts what it can, asks about the rest. Your estimator is teaching more than reviewing.
The questions narrow
Extraction is clean. Questions narrow to real feasibility calls. First drafts become worth editing rather than rewriting.
Routine comes back drafted
Routine RFQs arrive drafted and flagged. Your estimator spends their time on the hard ones and the ones worth winning.
You’re not buying a model that already knows manufacturing. You’re buying one that learns yours, and a team that stays until it has.
Between the order and the quote, there’s a box nobody has ever drawn.
The two ends of quoting are well understood. An RFQ arrives, a quote goes out. What happens in between isn’t on paper. Someone knows the part will be a problem before they can say why. Someone remembers a similar job three years ago. Someone knows this customer always changes the tolerance after award, and prices accordingly. None of it is in your ERP. It lives in two people, and it walks out at five o’clock.
Whatever the customer actually sent: the drawing, the spec, the revision, the email where the tolerance changed.
How your estimators actually judge a job.
Past quotes, process capability, what things really cost you.
The steps nobody ever wrote down. This is what we build.
A first pass with the risky parts marked.
Nothing reaches a customer unsigned. The agent drafts. The engineer decides.
By the end of a pilot the middle isn’t a black box in someone’s head. It’s written down, it runs, and your senior estimator can go on holiday.
Inside the workspace.
A normal RFQ tool at first glance. The agent inside handles the routine, asks when it isn’t sure, and gets faster as it learns your shop. Three views your estimator lives in.
The quote starts filled in
The agent reads the package and fills the quote skeleton: material, volumes, tolerances, delivery. Every value links to the line it came from, so checking is a glance, not a hunt.
It brings the precedent
It searches your past quotes and jobs for parts like this one, what they actually cost, and how they went. An unfamiliar standard or alloy, it researches and cites. Your shop’s judgment needed, it asks.
The last GJS-400 job needed a second heat treatment. Assume the same here?
The risky parts, up front
A feasibility view marks what makes the job hard, what’s missing, and what contradicts. Your estimator decides quote, decline, or ask, and the backlog stops being an inbox.
- Tolerances tighter than your process holds
- Contradictions between the drawing, the spec and the email
- Jobs that look like one you quoted before and lost, and why
When it doesn’t know, it says so.
The dangerous version of this technology fills every field with a confident number, and then you win a job you can’t make at that price. Ours separates what it read from what it inferred from what it’s unsure about, and puts the unsure pile in front of a person. Every value traces back to the line on the drawing it came from.
Five things move.
Every one of these is measurable in numbers you already track. That’s deliberate. We’d rather be judged on your scoreboard than introduce a new one.
Time per quote goes down
The extraction, the cross checking, the digging through old jobs: hours of it, on every RFQ, gone. Your estimator starts at the judgment call instead of at data entry.
Mistakes go down
A transposed dimension, a missed line item, a tolerance nobody checked. A quoting error is either a job you lose or one you win at a loss.
Cost per quote goes down
Same output, less senior engineering time in it. The direct saving, and the easiest to verify.
Throughput goes up
More RFQs answered per week without another estimator on payroll. The constraint stops being how many quotes your people can physically produce.
Revenue goes up
You quote work you’d otherwise decline, and you answer faster. Buyers expect a quote inside a day; a supplier who takes a week is often out before anyone reads the price.
The one to watch is throughput.
Time saved is easy to measure and easy to absorb into a normal week. Orders you couldn’t quote before are revenue that didn’t exist. If a pilot only proves time saved, it wasn’t worth running.
Questions, answered.
THE BACKLOG, QUOTED
Bring one drawing.
Half an hour. Send an RFQ you’ve already quoted, we run it live, and you compare it against what your estimator produced. If it doesn’t read your drawings, you find out on the call instead of after a purchase order.