Pulling your punches: The greatest leap in CNC CAM programming since punch cards

Posted on 15 Dec 2023 by The Manufacturer

With its AI-driven software solutions, CloudNC is aiming to help manufacturers raise productivity, shorten lead times and cope with industry’s widening skills gap. Jonny Williamson reports

It’s hard to overstate just how revolutionary the use of punch cards to control machine tools was. The technique marked a significant shift from manual control methods and laid the foundation for the development of modern Computer-Aided Manufacturing (CAM) systems. Before the arrival of punch cards in the 1950s, machining processes were labour intensive, error-prone and limited in complexity.

Punch cards provided precise and repeatable instructions for machine tools, with each hole representing a specific command. Almost overnight, machining processes became automated and repeatable. The advent of software-based CAM in the 1970s represented another major leap forward. It introduced the use of computers to generate and control toolpaths, unlocking a level of complexity and precision previously unreachable.

Further advancements have come in the half-century since. Yet, beyond the transition from 2D to 3D modelling, these have all been more incremental improvements than transformational shifts. As a result, Computer Numerical Control (CNC) programming has remained a largely repetitive, tedious and manual task.

Frustrated by this lack of progress, Theo Saville, CEO and Co-founder of CloudNC, has spent the past decade devising a better approach. A way to make CNC programming as intuitive and user-friendly as 3D printing, for manufacturing to foster rather than impede innovation, a way for companies to hire, train and upskill staff faster.

The result is CAM Assist, an AI-powered plug-in for Autodesk’s Fusion 360 software platform that has quickly garnered a reputation for being the biggest paradigm shift for the precision manufacturing sector in decades. I sat down with Theo to learn more.

Presentation CloudCN - Jonny article

What makes CAM Assist so revolutionary?

TS: CAM Assist is an AI co-pilot that works straight out of the box in your existing CAM package. It automatically generates professional 3-axis machining strategies in seconds at the click of a button. Research from our user community shows that it can get you 80% of the way there, with the remaining 20% more like reviewing the work done and filling in any blanks. This can save hundreds of production hours, completing tasks in seconds that currently take hours or even days.

This brings obvious advantages to manufacturers, including faster and more efficient programming without templates or macros, training up junior staff to become skilled machinists more quickly, and accelerating the estimation and quoting process. In short, CAM Assist helps you win more deals, complete jobs faster and create the capacity to take on more work. It’s a virtuous cycle.

You studied Manufacturing and Mechanical Engineering at the University of Warwick before joining the Metal Additive Manufacture Research team at Warwick Manufacturing Group. Do you see any shortcomings in the way CAM and CNC are taught?

Definitely. My biggest frustrations are around how time consuming and manual manufacturing is. I fell in love with the 3D printing process while at university. I had access to all these amazing machines covering FDM, SLS, stereolithography and more. I never read an instruction manual for any of them, they were just intuitive. I could design anything I could think of and go straight to manufacture in moments. Eventually, however, I needed to start making metal components and that’s when I started to encounter problems. One module involved designing and making a piece out of metal and introduced us to CNC machining.


The first time I was exposed to industrial CAM software, which shall remain nameless but remains very popular, I was shocked and not in a good way. It didn’t do any of the things I was expecting it to do. It didn’t suggest tools, feeds or speeds, nothing was automated and the interface was very clunky and unintuitive. It was just so manual – more manual than I thought possible. It was taking me 12 weeks to get something through the workshop when the term was only ten weeks.

I found the whole process exasperating, so I began asking questions. Why is this so hard? Why do I need to give a 2D drawing on paper to the manufacturing engineers in the workshop? Why are none of the students able to use these machines? Why does it take so long when most machines are idle all day? That was the genesis of CAM Assist. When I was at university, there was already talk of a skills gap. Now, we talk about a skills crisis.

Fewer people are entering industry every year than leaving or retiring. We aren’t going to solve this by simply finding more talent. And the software currently in use doesn’t solve the problem of nobody being there to drive it. We need software that drives itself. Software that enables people to go ten times further and faster for the same amount of effort.

That’s what CAM Assist is; software that makes it easier for people to use CNC machines and autocompletes as much of the task of making a component as possible.

Your frustrations with CAM software are, I imagine, fairly universal. So, why hasn’t someone addressed them already?

Honestly? Because accelerating precision manufacturing with software is really hard! Chris [Emery, Co-Founder and Chief Scientist, CloudNC] and I always knew it was going to be difficult, but the further we got into the problem, the more sub-problems we found.

Many of them so complicated that even the smartest algorithms couldn’t solve them quickly and required entirely new computer science to solve.

Rather naively, we thought we could crack it in six months. It ended up taking several years. It took so long because there are A; infinite combinations of solutions, and B; the physical process of machining that the software has to handle is incredibly challenging. We had to build up a huge code base of knowledge to solve all the disparate bits and pieces of the problem.

But I’m pleased to say that CloudNC launched the CAM Assist beta programme in January 2023, and in September, it became publicly available through the Autodesk App Store. You also operate a high-spec manufacturing facility in Chelmsford, Essex.

How important was having your own factory in the development of CAM Assist?

Crucial. We knew we needed somewhere to test our solutions and we figured that no factory would allow us to use their production machines for experiments. Building our own factory gave us our own testing ground, as well as allowing us to measure the impact of our software on day-to-day operations.

The factory allowed us to build software much faster, to test new capabilities and functionality and, if beneficial, to share that with our user community. This is one of the great benefits of the software engineering industry and one of the great tragedies of manufacturing and engineering.


If I solve a complicated problem once, I can share it with the entire world, either as a paid update, free patch or open source. Anyone can take my work and stand on my shoulders to solve the next problem. That process is replicated hundreds if not thousands of times a day. In manufacturing, the wheel has to be reinvented every single day. Knowledge is confined to one person, one factory, one company. It’s not distributed or shared. That needs to change and urgently.

In 2022, CloudNC announced Autodesk and Lockheed Martin as strategic partners. What do those relationships mean to you?

The visions of CloudNC and Autodesk are very closely aligned. Both companies want to leverage technologies like cloud, AI and machine learning to accelerate the design and make process. Achieving that means reaching as many people as possible. Fusion 360 has millions of users spread all around the world. Having access to that size of community is incredibly exciting.

Collaborating with Lockheed Martin helps us build software tailored for high-end industries such as aerospace and defence – complex components with complex requirements, made from complex materials. Gaining support from such giants of industry is a fantastic endorsement for what CloudNC has achieved so far and means we can keep moving ever-closer to our goal of end-to-end autonomous manufacturing.

How far off achieving that vision are you?

For CNC milling, about five years. For an end-to-end process, from quoting and estimation to process design, multi-machine programming and inspection, probably ten years. Not that long in the grand scheme of things. AI-powered tools like ChatGPT and Midjourney have changed the game. If something is easy and quick, we do it more often. I do so much graphic design work now that Midjourney can create all my images. Before Midjourney, I’d have to use a freelance marketplace like Fiverr, write a specification, hire someone, wait for them to complete the job and pay them.

To acquire the skills to do it myself would take too much time and money. Now, I enter some prompts and, within minutes, get something that is exactly what I need or can be refined and tweaked until it becomes perfect. The same is going to happen in manufacturing. It already is! The barriers to entry are constantly being lowered which means more and more people can bring their ideas to life. That’s unbelievably exciting, who knows what innovations it will lead to.


Earlier this year it was announced that CloudNC will supply precision manufactured components to the fusion energy industry, thanks to an agreement with the UK Atomic Energy Authority (UKAEA). The agreement will see CloudNC supply components that support the manufacture of diagnostic windows. These allow observers to view and monitor critical occurrences within complex equipment, such as reactions that take place inside a fusion machine.

The components will be manufactured at CloudNC’s factory in Chelmsford, Essex, and play a vital role in the development of future fusion power plants in the UK and the world. It’s the first time that CloudNC has manufactured components for the fusion industry. “These components are critical to the development of new research programmes and fusion power plants, and need to be made to the highest specifications,” said Samuel Nouvellon, Project Manager at UKAEA for the development and production of the diagnostic windows.

“Manufacturing components for the fusion energy sector presents unique challenges, as most commercially available materials and techniques cannot withstand close proximity to such fundamental reactions. “As a result, the parts must be manufactured to exacting standards, requiring advanced metrology equipment, as well as a meticulous focus on quality – both of which we provide,” commented Theo

For more stories like this, click here