In the fast-paced world of modern manufacturing, the integration of imaging and machine vision, powered by AI, is far from a futuristic concept - many would even go as far as to call it a necessity. From semiconductor fabrication to EV production, the demand for precision, efficiency, and automation is driving a manufacturing revolution. But as manufacturers look to leverage the power of vision AI, a key question arises: should they build their own solutions or buy them off-the-shelf?
According to Mark Priestley, Manufacturing Director at Robovision, this "build versus buy" dilemma is not just a technical consideration; it’s a strategic imperative that could impact a company's long-term success.
Mark Priestley, Manufacturing Director at Robovision
Priestley brings 27 years of manufacturing experience to this discussion, having worked as an automation engineer, robotics engineer, software developer, and programme manager, he has witnessed firsthand the consequences of this decision across numerous organisations.
The allure of the internal build
Many manufacturing organisations, particularly those with strong in-house technical teams, are drawn to the idea of building their own vision AI solutions. The perceived advantages are compelling. Control and Customisation, for example. Organisations will want to benefit from the ability to tailor a solution precisely to specific needs and integrate it seamlessly with existing processes.
There is also a belief that using existing resources and avoiding the upfront cost of commercial software can lead to long-term savings, while the opportunity to foster innovation and enhance the skills of in-house teams is certainly a highly sought after advantage.
Priestley says: “Organisations can make the mistake of almost acting like software development houses and trying to build their own software. Experience tells us that’s not the right approach for everyone.” A common pitfall in this case is underestimating the complexity and long-term implications that in-house development could bring.
The hidden costs and challenges
These pitfalls are not just the reserve of start-ups and smaller businesses, even the largest, most sophisticated organisations can face the same problems. Priestley elaborates: “Over the course of 10 or 20 years, that's created a massive problem for these companies, because they've got numerous systems, numerous devices and solutions that have been custom built by individuals within their organisation who are no longer there. These software platforms become unsupportable and they can't extend them anymore.”
One example Priestley cites is a global organisation that halted an audit after identifying tens of thousands of internal systems. This highlights a critical issue: “When they decide they're going to build their own things, it gets to a point where it can become completely unsustainable.”
Key challenges
Maintenance and updates can prove challenging, and keeping pace with rapidly evolving technologies requires a dedicated team and significant resources. “Technology moves at such a rate nowadays, you're probably talking about a lifespan, maybe six months,” Priestley explains, particularly in the realm of vision AI, where algorithm accuracy is constantly improving.
Custom-built systems can also often lack the flexibility to scale or be replicated across multiple production lines or facilities. “You build software and solutions with the intent to duplicate,” says Priestley, “and when you when you go down the road of building your own systems, the ability to be able to scale, the ability to be able to be agile, the ability to be able to duplicate those developments is close to zero.”
Then there is the time-to-market to consider. The development process can be lengthy, delaying the realisation of benefits. Priestley says: “Time is really critical here. Companies should focus on trying to receive the benefits as quickly as possible. I know of one organsiation that developed its own vision system, which took a number of years," Priestley explains. "Our product delivered the capability they needed out of the box, so they could have had the system up and running years ago."
Another challenge is the reliance on a few individuals for system maintenance. This presents a significant risk, especially as the workforce ages. “Imagine you're an organisation that has built homegrown systems, and the people who built them are now leaving your business,” Priestley warns.
Building software around existing processes can also actually perpetuate inefficiencies, rather than solving them. “In most cases, they don't know how to do it any other way. Of course, they start building inefficiencies into the solutions, the applications,” Priestley explains.
The advantages of off-the-shelf solutions
Off-the-shelf solutions can offer several compelling advantages. They can be implemented quickly, delivering immediate benefits and a faster return on investment. Subscription-based models can also provide financial transparency and ease of budgeting.
In addition, commercial vendors offer market-proven products designed to scale with business growth, with rigorous testing and ongoing support. Vendors also provide ongoing maintenance, updates, and expert support, to ensure the system remains current and effective. Says Priestley: “It will always come to the same conclusion that organisations get to a point where they can no longer accept the risk of in-house built systems, but by then they've already invested several years and a team of people to build these systems.”
Despite their advantages, commercial solutions aren't without potential drawbacks. The initial investment needs to be clearly thought out for those with limited capital resources. They may also require adjustments to existing workflows and processes, and organisations with very distinctive needs might find some compromise necessary.
Relying on external vendors creates a long-term relationship that can also carry some risk. Changes in vendor support, product roadmaps, or even business viability could impact system reliability.
The hybrid approach: best of both worlds
Priestley suggests a hybrid approach, leveraging off-the-shelf platforms that offer configuration flexibility. “You can take an off-the-shelf product that gives you the flexibility to build things in a controlled manner,” he explains. This allows manufacturers to customise solutions to their specific workflows while benefiting from the reliability and support of a commercial platform.
This approach acknowledges that manufacturers are experts in their core business, not software development. “A manufacturing environment and a manufacturing company are specialists in manufacturing,” Priestley asserts. By focusing on their strengths and leveraging external expertise, manufacturers can achieve optimal results.
Strategic considerations for the future
As the manufacturing landscape continues to evolve, the decision to build or buy vision AI solutions will be a critical strategic choice. By carefully weighing the benefits and challenges of each approach, and by considering the long-term implications, manufacturers can make informed decisions that drive innovation, efficiency, and sustainable growth.
"Sometimes they need to think about things more strategically and ensure they're looking after the long term interests of their businesses," states Priestley, and Robovision can help. “We try to establish an open conversation and have more of a partner-type attitude,” says Priestley.
“It's a collaborative effort between us and our customer base and our partners. So, we need to understand the challenges and the issues that they may be having, and then we work collectively to try and work out the best way forward to resolve and overcome those challenges. In some instances, we can upskill the clients and partners so they can go and do everything themselves, or in some instances, we'll help them do it. But the start of the process is understanding where the challenges are within the organisations and the steps that we need to take in order to overcome them.”
