Proof in the probe for AI-driven CAM

For those thinking AI in CAM is all about speed, CloudNC would probably agree. But those spending a bit of time at its Chelmsford factory quickly realise another story runs alongside: measurement. Lots of it. And not just as a tick-box exercise, but as the backbone of how the company builds trust in its software.

Setting the scene

CloudNC is the team behind CAM Assist, an AI that plugs directly into leading CAM systems. Manufacturers can fire it up inside their usual CAM environment where it gets to work analysing complex 3D geometry, generating machining strategies, and recommending speeds and feeds in seconds. No templates. No feature-based shortcuts. Just AI doing the heavy lifting.

“Programming CNC parts is essential, but it’s also time-consuming and, frankly, often repetitive,” states Dr Andy Cheadle, CTO. “CAM Assist is about removing that burden so programmers can focus on decisions that really matter.”

The numbers are compelling. With over 1,000 shops worldwide now using the software, CloudNC says programming time reductions of up to 80% are achievable. And with CAM Assist 2.0, launched late last year, things have moved on again.

“It’s no longer just about speed, it’s about interaction,” explains Andy. “We’ve built a workflow around ‘Configure, Assess, Prepare and Analyse’ so it becomes a dialogue between the user and the AI. You can set machine parameters, upload tool libraries and even configure the system from existing CAM files to mirror established processes.”

Then there’s the new Strategy Editor, introduced this year, which adds another layer of control. Users can review and tweak strategies before committing to toolpaths, moving operations between set ups or overriding decisions where necessary.

“Trust is everything with AI,” remarks Andy. “Strategy Editor gives users visibility and control. You can shape the output, not just accept it.”

Closing the loop

Here’s where things get interesting. Supporting all of this is CloudNC’s UK factory in Chelmsford, a fully operational machine shop that exists not to produce parts for customers, but to stress-test the software in the real world. Inside are a dozen or so high-end CNC machines from brands that include DMG Mori, Haas and Citizen.

“We built the factory to replicate the exact pain points our customers face,” he reveals. “It’s about closing the loop. We develop, test, measure and feed data back into the product.”

Walk into the quality and inspection room and there’s some serious kit in action. At the heart of it are two Hexagon Global S Chrome CMMs, built for high-speed, high-accuracy scanning. With vibration-reduction technology and optimised measurement paths, the CMMs make light work of complex geometries.

Out among the machines is a Hexagon Absolute Arm 85, offering flexible, shop-floor-friendly measurement with tactile probing and rapid 3D scanning. This portable and robust CMM allows engineers to capture data directly where it’s needed. And that’s the point.

“On-machine probing provides immediate feedback, but our selection of Hexagon fixed and portable CMMs let us go deeper. We’re dealing with parts that sometimes have extremely tight tolerances. You need confidence in what you’re measuring.”

Trust built in microns

This isn’t about verifying parts for the sake of it. It’s about validating the AI. The metrology is effectively checking that CAM Assist is working as anticipated.

It’s a continuous cycle: AI generates; machines execute; metrology verifies; and results inform the next iteration. In other words, the factory isn’t just a proving ground, it’s a learning engine. And that focus on real-world validation subsequently feeds into customer confidence.

Rapid ROI

Notably, CloudNC says achieving return on investment within six months is the norm, not the exception. At MACH 2026 earlier this year, the company supported this claim by offering a money-back guarantee for companies placing an on-stand order for CAM Assist if payback didn’t materialise in this timeframe.

“That offer came from what we see in Chelmsford every day,” reports Andy. “We’re not guessing. We’re measuring.”

And perhaps that’s the takeaway here. In an era where AI headlines tend to focus on speed and automation, CloudNC is quietly making a different point: none of it matters without proof.