Machine shops and CNC job shops can use AI to build custom machine shop management software for quotations and estimating, job cards with operation routing, machine loading, setup and cycle time capture, tooling and insert consumption, outsourced operations, inspection records, rework and job-wise costing. Pentoggle is an AI platform that generates production-ready software from a plain English description, which means a machine shop can build an application around its own machines, operations and customers instead of adapting to a system designed for a plant ten times its size.
Most machine shops in India already run Tally for accounting, some run Busy or Marg, and larger businesses may run SAP Business One. Those systems handle purchase, sales, GST and accounting well. This is not a proposal to replace them. Pentoggle builds the shop-floor application around them, covering the workflows they were never designed for.
Machine shops earn differently from manufacturers who own a product. A shop with its own product recovers development cost over volume. A job shop sells machine hours and skill, quoted per piece, against a drawing supplied by someone else. The rate is set in a negotiation you rarely win, so the margin is made after the order, in how efficiently the job runs. That is why a shop can be full of work all year and still not know which jobs paid.
Key takeaways
- Machine shop software is built around the job card and its operation routing, because that is the only record that connects a customer order to what actually happened on a machine.
- In low volume and high mix work, setup time is the real cost driver, and most shops record cycle time while ignoring setup entirely.
- Repeat orders are where the money is, and most shops re-solve them from scratch because the program, fixture and setup notes live in one operator's memory.
- Jobs disappear most often at outsourced operations such as heat treatment, grinding and plating.
- Quoted cost against actual cost, job by job, is the loop almost no shop closes.
Why machine shops are badly served by existing software
Nearly every machine shop in India runs Tally, and Tally does the books properly. It does not know that job 4471 is on VMC-3, that it has been waiting two days for heat treatment, or that the setup took four hours against the ninety minutes you quoted. That information lives on a whiteboard, in a supervisor's diary and in a WhatsApp group, and it evaporates the moment the job dispatches.
The alternative has been a packaged manufacturing ERP. Those are built around production orders against a bill of materials, with material planning at the centre. A job shop does not work that way. The material is often supplied by the customer, the bill of materials is a single part, the routing changes depending on which machine is free, and the same part number is quoted differently for two customers. Shops that buy these systems typically use the purchase and stores modules and abandon the production module within a year.
What is left is Excel, and Excel is genuinely good at estimating. It is bad at telling you where a job is right now.
Most shops are running some combination of the first two columns below.
What machine shops use today, and what they can build instead
| Whiteboard and Excel | Packaged manufacturing ERP | Application built with Pentoggle | |
|---|---|---|---|
| Job cards | Printed sheet, filed after dispatch | Production order against a BOM | Your operation routing, your machines, your terminology |
| Machine loading | Whiteboard, redrawn every morning | Finite scheduling almost nobody configures | Simple loading view your supervisor will actually keep updated |
| Setup versus cycle time | Cycle time estimated, setup ignored | Standard times assumed | Both captured separately, actual against estimate |
| Repeat orders | Rebuilt from memory | Order history without setup context | Part history with program version, fixture and last actual times |
| Outsourced operations | Challan book, tracked by phone | Subcontract module, rarely used | Challan chain with ageing and follow-up |
| Job-wise costing | Attempted annually, abandoned | Costing that assumes standard rates | Quoted against actual, per job, computed as it runs |
| Inspection records | Paper reports in a file | Quality module priced separately | First article and in-process records tied to the job card |
What a machine shop can build
Each of these can be built separately or combined. Most shops start with the job card and machine loading.
Enquiry and quotation register
Customer, drawing reference, material, estimated setup and cycle time, quoted rate and the revision history when the customer pushes back.
Job card with operation routing
Turning, milling, VMC, grinding, heat treatment and inspection, each with the machine, operator, planned time and actual time.
Machine loading view
What is running on each machine, what is queued behind it, and what is waiting for material or an outsourced operation.
Setup and cycle time capture
Recorded separately at the machine, so the shop learns what setups actually cost rather than what they were assumed to cost.
Part history
Every time a part has been made, with the program version, fixture used, tooling, actual times and any problem noted. The single highest value module for a shop doing repeat work.
Tooling and insert consumption
Inserts issued against jobs and machines, with consumption by part, which turns tooling from an untracked overhead into a per-piece cost.
Outsourced operation tracking
Challans out to heat treatment, plating or grinding, with ageing and expected return, so a job stops disappearing for a week.
Inspection records
First article, in-process and final inspection tied to the job card, with the report retrievable when the customer asks six months later.
Job-wise costing
Quoted against actual, using recorded machine hours, tooling, outsourced operations, scrap and rework.
Setup is the cost, and almost nobody measures it
A shop running batches of twenty and fifty pieces spends a large share of its available machine hours not cutting metal. The machine is stopped while a fixture is changed, a program is loaded, offsets are set and a first piece is approved. Cycle time gets estimated carefully because it is easy to estimate. Setup gets an estimate pulled from experience, and that estimate is rarely checked against what actually happened.
This matters because the two behave differently. Cycle time is recovered per piece and scales with quantity. Setup is a fixed cost absorbed across the batch, so on a batch of twenty it can exceed the entire cutting cost. A shop that quotes on cycle time alone will systematically underprice small batches and overprice large ones, and will conclude that small batches are unprofitable in general rather than that its estimate was wrong.
Capturing setup separately at the machine, even as a start and stop entry on a tablet, gives a shop the one piece of data it has never had. After a few months the estimate becomes evidence, and the quotation improves without anyone changing the rate.
Repeat orders should get easier, and usually do not
The second time a part comes in, the shop has already solved it. The program exists somewhere, the fixture is on a rack, the tooling is known and somebody remembers the sequence that worked. In most shops that knowledge sits with one operator or one programmer. When they are on leave, or when they leave, the part is re-solved from the drawing.
A part history record fixes this without changing how anyone works. When a job is closed, the program version, fixture location, tooling list, actual setup and cycle times, and any problem encountered are attached to the part. The next time the order comes in, the quotation is built from evidence rather than estimate, and the setup on the floor starts from what worked last time.
This is also the argument for accepting a repeat order at a rate you would refuse for a new part. If the shop can see that the part ran at eighty percent of estimated setup last time, the negotiation changes.
Jobs disappear at the outsourced operation
Heat treatment, grinding, plating and special processes usually go out of the shop, and that is where jobs vanish. The material is on someone else's floor, the follow-up happens by phone, and nobody notices a delay until a customer asks. If your customer supplied the material, that movement also carries documentation obligations, and the challan chain has to hold together from their inward challan through to your return.
An application built for your shop tracks the challan out, the expected return date, the ageing and the value sitting outside, and shows those jobs on the same loading view as the machines. A job at the heat treater and a job on VMC-3 are both work in progress, and a shop that sees only one of them is planning with half the picture.
For the documentation side of customer-supplied material, see the job work software guide.
Drawings, revisions and who is holding the old one
A drawing changes after quotation. Revision C becomes Revision D while material is already cut, and the shop now has three questions at once. Which jobs have already started on the old revision, and which can still switch before machining begins. Who on the floor is holding a printout that is no longer valid.
The answer in most shops is a phone call to the customer and a walk around the shop, and the cost of getting it wrong is a batch scrapped or a delivery rejected at inspection. It is the kind of problem that happens a few times a year and is expensive every time.
Handling it does not require a document management system. It requires the drawing number and current revision to sit on the job card, the file itself to be attached and viewable at the machine, a revision history that shows what changed and when, old revisions retained rather than deleted, and where a customer requires it, a record of who acknowledged the change. Once revision is a field on the job card rather than a note on a printout, the question of which jobs are on which revision answers itself.
Why building this is now practical
A shop with twelve machines has never been able to justify custom software. A development team, a specification and a six month build were never going to be repaid by a business selling machine hours at rates the customer sets.
That has changed. With Pentoggle you describe how your shop runs, including your machines, your operation sequence, your customers and the way you quote, and get a working application. When you add a machine, take on a customer with a different inspection format, or start tracking tooling by part, you describe the change and the application updates. Shops usually begin with the job card and loading view and add costing once the time data has accumulated for a few months, which is the right order because costing without actual times is just the estimate again.
Why machine shops choose Pentoggle
Built around your operations
Your machines, your routing, your inspection points, your job card format.
Works alongside Tally
Pentoggle handles the floor, the job cards and the costing. Your accounting stays where your CA already works.
Built for Indian documentation
GST invoicing with HSN codes, delivery challans, e-way bill workflows and challan records for customer-supplied material.
Usable by the people on the floor
Simple entry on a phone or tablet at the machine, because a system your supervisor will not update is worse than the whiteboard.
Changes in days
A new customer requirement does not become a three month project.
The one number that runs a machine shop
Machine utilisation, split into cutting hours and setup hours.
Total utilisation on its own is a comforting number and a misleading one, because a machine occupied all day with three setups and two hours of cutting looks identical to one running a full shift. The split is what tells you something. It shows whether your shop is short of work, short of planning, or losing its capacity to batch sizes that are too small for the way you quote.
Track it per machine, weekly. If cutting hours are low and setup hours are high, the answer is usually sequencing similar jobs together rather than buying another machine.
Ready to build machine shop software?
A machine shop rarely runs out of work. It runs out of visibility.