While modern manufacturing has deployed robotics and automation at scale, many of these deployments have not been complete successes. According to the International Federation of Robotics, over 500,000 industrial robots are installed every year, with the total number of robots deployed globally now nearing 5 million.
A 2025 Deloitte study found that this is likely to increase, with 41% of respondents intending to invest further in factory automation hardware. However, most factories do not see the true RoI – instead, they are filled with underutilized robots, suboptimal hybrid human-robot lines, and large manual rework or exception-handling stations.
All of these are consequences of a shift that has yet to take place across the industry, from traditional Design for Manufacturing and Assembly (DFMA) principles to Design for Automated Manufacturing and Assembly (DFAMA); to make this shift requires changes in both product and process design philosophy.
The Physics of Automation
While robots can significantly increase productivity through 24x7 operation and increased throughput, there are fundamental constraints on what they can do.
First, robots demand deterministic processes, with predictable and repeatable steps and states. Unlike humans, who can be adaptive and deal with ambiguity, robots need perfect clarity and carefully controlled environments; processes that depend on “feel” cannot be automated reliably.
Second, robots have limited dexterity and often struggle with deformable or flexible parts. Parts that humans can handle easily and instinctively, such as cables or parts that can be randomly oriented, cause robots to fail; designing for automation requires parts that are easy to grip.
Third, robots have sensing limitations. Vision systems that underpin modern robotics tend to struggle with transparent or reflective surfaces, or when parts have random orientation or asymmetry and need to be reoriented before assembly – tasks that humans can easily complete based on judgment and experience.
Finally, robots are most efficient when part counts are smaller and parts are more consistent. More parts imply more robotic actions, thereby increasing the number of failure points. Multiple actions, such as mixed fastening strategies (for example, screws, adhesives, and clips), complicate automation sequencing and increase the risk of exceptions or errors that compound.
Traditional DFMA approaches, such as the Boothroyd-Dewhurst and Lucas methods, were developed before the modern era of automation. As such, they optimized for adaptability, a core trait that distinguishes humans from machines. Designing for automated manufacturing and assembly (“DFAMA”) requires repeatability and consistency – typically at the expense of adaptability and flexibility.
As such, retrofitting automation solutions on existing processes and product designs is a recipe for failure. Most product designs today are fundamentally incompatible with automation and need significant redesign. Attempting to retrofit determinism onto processes designed for human flexibility results in mixed processes that are suboptimal for both humans and robots.
Building automation that works
Design, manufacturing, and operations teams that are leading the shift towards DFAMA follow a few key principles.
- Increased collaboration: Product designers, automation engineers, and factory operators collaborate early in the design process and continue working together throughout manufacturing. Often, the best teams collaborate closely with their leading suppliers to involve them in early design decisions that affect automation compatibility. A simple example of a use case that requires cross-team design collaboration is defining appropriate tolerance sensitivities for interlocking parts – not too loose to require human judgment, but not too tight to cause exceptions due to extreme precision requirements that are beyond robotic capabilities.
- Standardization and simplification: Designs and processes need to be set up to be as standardized as possible. Unified interfaces, consistent fastening methods and regular geometries are all automation-friendly. Other forms of simplification, such as reducing the number of parts used, clearly marking correct orientations and alignment features, and reducing the number of process steps, all improve yield from automated manufacturing by reducing the number of failure points and the potential for compounding errors. Some traditional DFMA concepts that eliminate process complexity still retain utility here, but need to be used carefully.
- Virtual validation: With advances in CAD, CAE, and CAM technology, designs and processes can now be virtually tested and validated prior to deployment. All key workflows can be simulated to test key metrics – throughput, yield, rejection rates, number of exceptions, and other such parameters. Consequently, parts and processes can be clearly set up for successful automation before making costly and time-consuming automation decisions.
The economics of automation
The success or failure of automation on the factory floor is largely determined before the first robot is purchased and installed. A successful ecosystem – involving collaboration across teams, parts and processes that are clearly designed for automation and rigorous testing to confirm the efficacy of automation solutions before installation – significantly increases the RoI from automation.
Automation solutions typically have high upfront CapEx costs and often require effort to integrate with existing systems and processes. However, these costs can be paid off over time through improvements in throughput and yield as well as reductions in labor costs.
Consequently, clarity around what automation (if any) can be effectively deployed and what changes need to be made to enable its long-term success is crucial, with significant economic and strategic implications for capital allocation, long-term cost structures, and supply chain strategies.
The best examples of successful DFAMA have been in automotive manufacturing, where over 1 million robots drive effective automation, underpinned by predictable and repeatable processes that use standardized, redesigned parts.
The way forward
Vikram Venkat, principal at Cota Capital.
Companies that nail the shift from ease of assembly to deterministic, machine-executable assembly will have a competitive advantage in an automation-first world; to do so will require a shift to DFAMA philosophies.
