Because manufacturers are constantly pressured to increase productivity, deliver quality products, delight customers and achieve greater profitability and sustainability, they are turning to new technologies such as artificial intelligence. To remain competitive, the time for such digital transformation is now.
However, barriers continue to exist and technology is not always being used to its full potential, as we learn in a recent article, “How to Outsmart AI Business Challenges” in SCM Now Magazine from the Association for Supply Chain Management.
About Artificial Intelligence
In his white paper, “No Matter What You Call It: Industry 4.0 Means Manufacturing in Transition”, QAD’s Glenn Graney, director for the Industrial and High Tech markets, reflects on artificial intelligence:
“Artificial Intelligence (AI) is already transitioning from an academic exercise to an impactful business proposition. The practical adoption of AI through machine learning results directly from enhanced connectivity, smarter sensors, advanced analytics and super scalable systems.”
While we don’t think that AI is a dirty word, we know that not all manufacturers have discovered the solutions promised by AI. Companies need to build the proper foundation to enjoy the explosive growth of AI, machine learning and similar technologies. Among other things, it takes best practices, simplified IT architecture, well trained teams and a proper IT investment – in other words, money, patience, planning and force of will.
Barriers (and Solutions) to Implementing Artificial Intelligence
No technology advancement occurs unscathed. Manufacturers face four main challenges from achieving the promise of artificial intelligence. Below are some potential solutions:
Fear of Missing Out (FOMO)
To avoid the dreaded FOMO, rather than rushing to act on the hype surrounding AI, manufacturers need to start projects with a clear strategy. As in all things, a well-designed plan is critical and should start with building a portfolio of use cases that demonstrate the technology’s value, identifying deliverables with deadlines. The plan should include route optimization, sales forecasting, product categorization, safety stock calculation, supplier management and warehouse management. It’s critical to clearly outline the value proposition and KPIs to be measured along with goals and objectives. Estimate ROI based on project-related costs.
Deployment Difficulty
Deployment is eased when manufacturers choose packaged applications from among the following three types, and it’s important to build in change management plans for better user adoption.
- Focused solutions or packaged applications with pre-built AI models which only require the input data.
- Embedded AI solutions from common applications, such as advanced planning, transportation management and warehouse management systems.
- Custom solutions from open-source platforms, frameworks and application programming interfaces which afford organizations the means to build custom AI models.
Data Readiness
Focus on data quality and relevance over volume and work with the best data first. Many AI-embedded supply chain solutions use smaller data sets to optimize their AI capabilities. Begin with the most relevant data first. Sometimes in machine learning projects for product categorization and clustering, companies decide to throw in many data sources without ensuring their relevance. A lack of data quality can lead to project failure. Moreover, data preparation is a time-consuming process that requires diligence.
A Lack of Talent
Don’t forget training to fill AI roles, as a lack of appropriate skill sets is a major obstacle to AI deployment. Training is essential, whether it’s in-house, online or offered by graduate programs at local universities. And training requires a budget line item. In addition, companies can find independent contractors and IT service providers to fill talent gaps.
“AI is becoming a proactive element to advanced manufacturing, product lifecycle management and enterprise asset management. Its proactivity is based on context and what it has learned, not simply based on established metrics such as performing preventative maintenance based on accumulated run time or placing an order based on a predefined reorder point.” – Glenn Graney, Director of Industrial and High Tech Markets
To learn more about ways to overcome obstacles to achieving efficiencies and competitiveness with AI, check out the full article in SCM Now Magazine.
Very interesting article! It super interesting and my favorite part is “The Barriers (and Solutions) to Implementing Artificial Intelligence”- super insightful!