Strategic Guidance for Smart Manufacturing: Artificial Intelligence Apps

The Oracle Cloud Platform now offers the new Oracle Adaptive Intelligent Applications for Manufacturing, which include:

• Pattern and correlation analysis: Matches data from the plant floor such as employees; equipment; processes; supply, and management of business outcomes, such cost, cycle time, quality, scrap, rework, yield, and returns, finding correlations and causes.
• Genealogy and traceability analysis: When a manufacturer needs to recall a product, this solution helps find the products, services, and customers specifically affected to limit the impact.
• Predictive analysis: Uses the correlation analysis to forecast future business outcomes and alerts users if any will likely drop to help prevent unplanned downtime.

Artificial Intelligence (AI) has come to manufacturing from a variety of different types of vendors and applications, and Oracle has now entered the fray. AI provides a device or software program with the ability to interpret complex data, including images, video, text, and speech or other sounds, and acts on that interpretation to achieve a goal. Oracle’s new apps provide examples of this, as do the many predictive analytics apps already on the market.

Other examples of AI-related vendors, products, and apps in manufacturing include FogHorn Systems’ Lightning EdgeML, Predii’s Natural Language Processing (NLP), Movidius’ Vision Processing Units (VPUs), PTC’s predictive analytics, Alibaba Cloud’s ET Industrial Brain, Landing.AI’s automated quality control and predictive maintenance, IBM’s Watson, SAP Leonardo Machine Learning Foundation, and Amazon Web Services (AWS). These various vendors all currently have commercial offerings with AI and machine learning capabilities.

AI vendors face enormous opportunities but even bigger challenges within the manufacturing market. Many could help make substantial improvements to efficiency and productivity across the sector, but most end users lack the ability to implement and use the technology. This means that AI vendors must:

• Ingest data any way possible: Several of the vendors listed above have their own Application Enablement Platforms (AEPs) or partners with AEPs. Unfortunately, in manufacturing, that does not necessarily mean they have access to the data they need because oftentimes the end users do not have this access. AI vendors need Operational Technology (OT) protocol translators or partners who specialize in data extraction. They need to adapt to each client’s situation; each will have unique requirements.
• Solve the core problems in manufacturing: This means lowering maintenance costs, of course, but also reducing time to market, increasing efficiency and productivity and improving quality control. Generative Adversarial Networks (GANs) will generate designs and evaluate new products, and machine learning in computer vision can catch anomalies on the production line.
• Work on an open-source framework: Open-source software frameworks receive support and regular updates from a large community of developers, making them far more likely to remain stable and scalable. This carries importance for vendors selling their AI as a service and whose clients will rely on the solution.
• Provide consulting for setup, data ingestion, data curation, and results delivery: Most end users in the sector lack the necessary expertise to get the most out of any AI-based app. Even if they have an army of computer scientists, they often require instruction on how to get started. To help clients reach Return on Investment (ROI), offer consulting on setup, data ingestion, data curation and communicating insights to the right people. As stated above, data ingestion may require bringing in other partners. If the clients lack computer scientists altogether, a managed service option might provide another revenue stream and lead to more successful deployments.

Following this strategic guidance should help AI vendors to better prepare for future opportunities in smart manufacturing.

For more insights and perspectives on manufacturing, please check out ABI Research’s Smart Manufacturing service.