The need for Artificial Intelligence : The evolution of AI has been undulating but today it has matured considerably, owing to rapid growth in computing power and the need for processing massive data, which is being generated at a rate never experienced before. This need has given rise to the growth of algorithms that can quickly extract meaning from data, draw patterns, and derive functional relationship. Manufacturing firms are not far behind in embracing the power of AI and have been front-runners in the adoption of technology. With Google already testing its self-driving cars and trucks, innovative companies like Audi and Volvo are not far behind in making plans to launch models with advanced self-driving capabilities this year. Then we have companies like GE and Siemens who have adopted AI to optimize power plant performance achieving significant results.
Next, let us check an example to understand the potential benefits of AI.
Artificial Intelligence – a practical application : The supplier selection process is critical for any firm. Today, when the business ecosystem spans across borders and time zones, the supplier spread is also getting extended to support low-cost global operations. For most firms, suppliers are selected on four main criteria – quality, cost, service, and delivery time. However, the decision-making criteria varies according to industries and at times for a firm. For instance, Engineering Procurement Construction (EPC) firms often have project-based requirement that may call for custom design changes. Thus, flexibility at the suppliers’ end to accommodate custom requirements holds relevance for such firms. Same holds true for suppliers that are part of aviation ecosystem, however in aviation industry, “Individual design changes happen at a very late stage”, and suppliers are expected to accommodate the changes without affecting the lead time. Likewise, there are more scenarios which further increase the complexity of the selection process.
AI methods such as Machine Learning (ML) can be explored to handle this problem. An ML model can correctly filter out suppliers who do not meet the requirements, and then rank them. The criteria variables are fed into the ML model to train it. Based on the training data, the model is refined and deployed. The model continuously learns on the data to improve its prediction accuracy. The model’s prediction is as good as data hence the data quality is critical for ML models to function. The identification of the algorithm is a tricky process since the same problem may have different approaches based on an understanding of modelers. However, it is important to understand not only the dataset (size, quality) but also available computation time to choose the right algorithm.
The road ahead : Manufactures in every domain will find AI relevant. Within manufacturing firms, it is the automotive and assembly that leads the adoption and there is still time for technology to become mainstream. However, firms who have proactively invested in AI have seen higher profit margins such as GE, Toyota, Tesla, and BMW.
To conclude, AI caught significant attention about six decades ago, thanks to Alan Turing’s work. It could never find its place in the technology mainstream, and progress has been limited to research, concepts, and few prototypes. With the exponential growth in digital critical blocks such as cloud, storage, analytics, and CPU/GPU processing power, this is the best time for organizations to start considering AI.