Closing the gap between design and manufacturing

Manufacturers have talked about design for manufacturability for decades, yet the same friction remains familiar: engineering completes a design, production reviews it, and problems emerge when there is little room left to change geometry, materials, tolerances or process assumptions. For Asif Moghal, Director, D&M Market & Industry Development at Autodesk, the issue is less about the capability of individual tools than the way work still moves between teams.

“The problem isn’t that engineers do not care about manufacturability, or manufacturers do not care about engineering,” Moghal says. “Manufacturing knowledge arrives after the most significant design decisions have been made. Manufacturability is reviewed at a point when the organization has the most information but the least freedom to act on it.”

That pattern has survived even as companies have digitized design, CAM, quality and supply chain functions. What has often lagged is the connective layer between them. Sophisticated systems can still sit inside an operating model built around sequential handoffs, leaving teams to exchange information after decisions have already hardened.

Move manufacturability upstream

Bringing manufacturing into the process earlier changes more than the timing of a review. It changes the relationship between the people designing a product and those responsible for making it. Rather than production being asked whether something can be manufactured, both sides can explore how it should be designed and produced while meaningful options remain open.

Moghal has seen manufacturers place engineers on the shop floor to improve awareness, but he questions how well that approach scales across an organization. The bigger opportunity, he argues, is to embed manufacturing knowledge into the development process so that collaboration does not depend on individuals being physically co-located.

“The relationship moves from observation and correction to collaboration,” he says. “Manufacturing is not just asked, ‘Can you make this?’ They get brought into the discussion of what is the best way to design and make it. The objective shouldn’t be to constrain creativity earlier. It should be to give better information upfront to inform the creative process.”

That becomes increasingly important because design decisions rarely have a single downstream consequence. Geometry may affect machining strategy, while materials can influence tooling and lead times. Tolerances and component count can also change production complexity and cost, while choices made in engineering may have implications for sustainability or supplier availability.

For Moghal, the goal is not to identify one decision that matters most, but to improve the quality of multiple decisions by exposing their consequences earlier. A complex geometry, for example, may be appropriate for additive manufacturing but unnecessarily expensive to machine. A material selected for performance may create an entirely different set of issues around energy use, recyclability or sourcing.

The argument is becoming more urgent as manufacturers operate with less spare capacity. Larger organizations may still be able to absorb inefficient processes with extra time and resources, but many smaller and midsize manufacturers cannot. The pressure is therefore growing to use technology not simply to digitize existing tasks, but to shorten development cycles and make better use of the expertise already available within the business.

Connect decisions, not just files

Autodesk positions Fusion, its design and manufacturing platform, as part of a broader move away from separate point solutions and sequential workflows. “What we need to make sure is that it’s not just providing multiple tools under one brand,” Moghal says. “It’s around connecting the data between these different systems, taking the data from engineering and exposing it to manufacturing and vice versa, but helping people make better decisions in context.”

A digital thread can still amount to little more than files moving between departments. A CAD model may contain far more information than a production engineer needs, while the critical detail for a particular decision could be a tolerance, material, revision status or tooling assumption.

For manufacturers, the challenge is deciding what information people need at each stage of the product lifecycle, rather than assume that moving more data between systems will solve the problem. “A digital thread is not just a collection of files moving between different departments,” Moghal says. “It’s the key information that must flow between the departments to help them make the best decisions around how they make the product. A file tells you what a product looks like, versus a digital thread, which tells you what was decided, why, and the decisions it affects.”

Simulation extends that idea by allowing more of the downstream impact to be explored before physical changes are committed. Traditional engineering simulation focused heavily on product performance, but manufacturers can now model the factory and its processes to examine how production is likely to behave and where bottlenecks could emerge.

At the more advanced end, Moghal describes the digital twin as a way to compare expected behavior with real-world performance. A factory model based on actual machine data can show how production should operate, while live results reveal where reality diverges from the model. The difference can then become a source of improvement rather than simply a performance problem.

Generative design pushes manufacturability further upstream again. Instead of creating a geometry and checking whether it can be made, manufacturing methods and constraints can become inputs to the design exploration itself. That allows software to explore alternatives at a scale no individual engineer could reasonably match, while keeping human judgment in the loop.

“Manufacturability is more than just checking the answer,” Moghal says. “It’s helping define the solution in the first place. It doesn’t remove engineering judgment, but it helps us explore at a scale and a speed that we couldn’t literally do as a human being.”

Build the feedback loop

The same principle applies once products reach the factory. Production inevitably uncovers information that was not obvious during design, whether that involves machining difficulty, quality variation, supplier feedback or a better way to produce a component. The challenge is turning those discoveries into reusable engineering knowledge rather than allowing them to remain buried in emails, corrective action reports or local workarounds.

Too much production knowledge is still lost once the immediate problem has been fixed, leaving the same lessons to be rediscovered later. “A closed loop would be able to tap into the standard tools that you use, like Teams, Outlook, email, Explorer, SharePoint, anywhere that some feedback is captured,” Moghal says. “If you could convert production evidence into timely, actionable learnings, every production cycle should theoretically improve. Not just fix the current job and then forget that we’ve got that opportunity to fix.”

AI could accelerate that shift, although its role still needs clearly defined limits. It can automate repetitive work, make specialist knowledge easier to access and help engineers explore alternatives, particularly as the problems they face become more complex.

The risk is confusing something that looks plausible with something that has been properly engineered. AI may generate a convincing image, model or recommendation, but manufacturing still requires somebody to judge whether the result is safe, practical and appropriate.

“It can expand engineering capacity and capability, but it doesn’t remove the need for engineering accountability,” Moghal says. “It’s not a tool to replace an engineer. I think it’s a tool to enhance the capabilities of an engineer and remove the need for that person to start from zero.”

Better software alone will not eliminate the old handoff model if organizations continue to make decisions in isolation. The bigger opportunity is to connect knowledge across product development and production so that manufacturing consequences become visible while engineers still have the freedom to respond to them.

For companies under pressure to reduce costs and accelerate development while making better use of scarce expertise, that could become an important source of competitive advantage. Technology can expose more information and explore more options, but its value ultimately depends on whether manufacturers can turn that knowledge into better decisions before the opportunity to change has passed.