Most M&D companies are looking for new ways to gain a competitive edge in their given fields. Throttled by their existing legacy systems and manual processes, these organizations are turning to advanced technologies like AI and machine learning for help. Unfortunately, those that haven’t already begun investing in Industry 4.0 automation and technologies still have some hurdles to jump before they can begin leveraging the power of artificial intelligence. Put simply, any company that’s still using a legacy green screen version of an enterprise resource planning (ERP) solution – and yes, there are plenty of them still in use out there – will have a difficult time implementing modern technologies. Even if the AI initiatives are being driven by senior executives and leaders, there are still some steps that have to be taken before M&D companies can successfully adopt AI, machine learning, and modern data platforms. Here are five roadblocks that must be addressed before M&D can truly realize the benefits of these innovations:

• Integration with existing systems. Integration of AI technologies with existing legacy systems is a huge hurdle for M&D companies that have long-established processes and systems that aren’t compatible with the latest AI technologies. To help ensure minimal operational disruption, the integration process must be meticulously planned out, the current IT infrastructure evaluated, and the compatibility of AI solutions carefully assessed. Also, a gradual, phased approach is often more successful than a complete overhaul. Depending on the age and condition of manufacturing equipment, integration of sensors and IoT devices can aid in the efficacy of AI adoption and result sets.
• Data management and quality. Artificial intelligence systems require high-quality, relevant data to function effectively, which means M&D companies must assess their current data collection and storage practices before jumping in. This is especially critical if the current data is siloed, outdated, or formatted in a way that’s incompatible with AI applications. Establishing robust data governance, ensuring data quality, and developing efficient data pipelines are essential steps, along with training staff on the importance of data accuracy and consistency. In many cases, pre-work (e.g., constructing a modern data management platform) will be needed before a proper AI strategy or adoption can be executed.
• Talent and skills gap. Successful implementation of AI requires a workforce that’s skilled in modern data architectures, languages, and platform technologies. If there’s a significant skills gap in current employees’ AI and data analytics capabilities, M&D companies should invest in training and development programs needed to upskill their workforce – or work with outside professional services firms to augment their capabilities. Additionally, organizations may need to hire and attract new talent with specialized skills in AI, machine learning, and data science to maintain these platforms once the technology is up and running.
• Cost and ROI concerns. Investing in AI can be costly, and it’s crucial for companies to have a clear understanding of the return on investment (ROI). This includes the costs of software and hardware, integration, training, and ongoing maintenance. Manufacturers and distributors should also develop a clear business case for AI adoption, outlining the expected benefits like improved efficiency, reduced downtime, and enhanced decision-making capabilities. Finally, they need to set realistic expectations and understand that the ROI from AI adoption may be realized over a longer term.
• Cybersecurity and privacy. Artificial intelligence systems often process and store sensitive data, making them targets for cyberattacks. Companies must ensure robust cybersecurity measures are in place to protect their data and AI systems. This includes regular security audits, employee training on cybersecurity best practices, and implementing advanced security protocols. Additionally, compliance with data privacy regulations, such as GDPR or CCPA, must be ensured to avoid legal complications. Steps should also be taken to ensure sensitive PHI, PII, or customer data isn’t uploaded into the AI large language models (LLMs) that exist in the public domain. Private and/or enterprise versions of these LLMs also exist, such as Open AI’s ChatGPT Enterprise, which can exist within the four walls of an organization.

The successful adoption of AI and modern data analytics in M&D requires careful planning; investment in technology and people; and a commitment to continuous improvement and adaptation. A good starting point for companies that want to begin the journey is to start with an AI-readiness assessment. CohnReznick uses this assessment as an opportunity to determine exactly where the organization stands from a maturity perspective as it relates to people, processes, and technology. Then, we factor in the outcomes that the company wants to achieve with AI and what the potential use cases are (e.g., back-office automation, quality assurance, assembly line automation, etc.). From there, a roadmap is developed that companies can use to adopt AI and other modern data platforms.