Developing a Strategic Approach to Production-Enhancing AI

artificial intelligence (AI) to substantially transform their businesses. The initial applications of generative AI (genAI) in industry are focused on increasing bottom line efficiency and productivity, cutting costs and improving product quality.

While these results sound promising, organizations must understand that building effective genAI frameworks and models requires a foundation of reliable data on which to base actionable suggestions, and there are no silver bullet solutions. Achieving success with genAI mandates a strategic approach to organizational readiness.

GenAI represents the next evolution in operational machine learning, enabling self-learning based on patterns in existing data. This technology brings the concept of the augmented engineer to life by suggesting solutions, answering questions and explaining problem-solving methods. Additionally, it accelerates the integration of human expertise with advanced data analytics.

With 2.1 million jobs in U.S. manufacturing projected to go unfilled by 2030 (based on a Deloitte prediction), companies will need to increasingly rely on AI to fill the void. AI enhances human capacity to address emerging challenges with insights derived from operational data.

Assessing organizational readiness

To assess readiness and augment process data analysis with genAI, organizations must first examine their data quality. High-quality data is essential for genAI effectiveness. A key aspect of this quality is connected to its relevance to the specific problems a team is working on. To fulfill these requirements, users need the knowledge and ability to prepare their data effectively (as shown in Figure 1). After all, technology’s output is only as good as the quality of the data. As the saying goes: garbage in equals garbage out.

Figure 1: Advanced analytics AI assistants are not a replacement for experienced application users, but they can help cleanse data, conduct analysis, and drastically increase speed to insight. Image courtesy of Seeq

GenAI users in industry need to be data literate, understand AI and machine learning tactics, and be able to interact with AI-powered tools. They must also be able to develop and maintain genAI solutions with an understanding of the process and business teams the solutions are targeted toward. Therefore, adaptability and continuous learning are crucial as AI technologies evolve.

One of the greatest barriers to this kind of ongoing technical training is a lack of time, especially in sustained blocks, to engage in formal training courses. Fortunately, AI can help with this. The most forward-thinking manufacturers are using AI to provide industry-relevant, just-in-time learning on demand, enabling teams to—as one Seeq AI Assistant user said—“complete tasks eight times faster and without interrupting other teams for help."

A key benefit of AI here is its ability to consolidate training and reference material, providing responses tailored to the specific equipment and process of interest.

Dustin Johnson, chief technology officer, Seeq.

Another challenge many manufacturers face is employee mistrust of AI based on early issues regarding hallucinations and inaccuracies. Businesses can address these concerns by holding targeted training sessions to demonstrate how trailblazers within the company are achieving success using AI. These sessions provide an opportunity for project leaders to share how rapid developments in prompt engineering and the retrieval of relevant documentation drives exponential improvements in accuracy. This increases confidence in the technology by providing the ability to trace results to verify the data and input models.

For industrial organizations, this means providing users with easy access to data so they can validate AI results, explore monitoring alerts, review diagnostics and evaluate their own hunches. For example, the Seeq advanced analytics and AI platform provides data contextualization capabilities to help users better determine the relevance of their results, enabling them to verify the data used to make important decisions.

Dustin Johnson is the chief technology officer at Seeq.