How to Get Started With AI to Demonstrate Business Value In Months
Part 2: AI-Driven Applications Must Capture and Combine Human Knowledge With Data for Digitalization to Succeed
In the previous post in this series, we looked at the approach many large industrial companies are taking for successful digitalization. They are assigning a CDO to assemble a healthy ecosystem that consists of an internal team, a systems integrator to augment their team, and a stack of best-of-breed digital technologies. In this post, we’ll examine what an effective digital technology might look like.
The AI-enabled software selected by a company should be open in nature to prevent the company from being locked into a technology that is custom-coded by the vendor and requires relying on that vendor for any adjustments or changes. There are two keys to successfully implementing AI-driven applications or solutions. First, CDOs must involve the experts that run those operations from the very beginning. Second, they must remember that AI is not a data science project, but instead it is a project that must capture human expertise and combine that with the most relevant data in order to appropriately train the AI algorithms.
One example of AI in action in an industrial setting is in the crude oil refinery area of any oil and gas company. The downstream process of refining crude oil into a finished product involves many potential risks, from equipment failures to unplanned downtime. Crude engineers at oil companies gather knowledge about the refining process — the chemical composition of crude oil, how to treat the oil to avoid corrosion, etc. However, not all crude engineers are experts in hundreds of crude types and when these subject-matter experts leave, the company loses the knowledge they possess and struggle to share that information with the broader engineering team.
When engineers don’t have the knowledge necessary to mitigate risks associated with different types of crude, the organization risks hundreds of millions of dollars associated with production loss and equipment failure. In fact, crude corrosion across all oil and gas companies globally, cause $15 Billion dollars industry loss. However, the implementation of AI-enabled technology, such as Maana’s Computational Knowledge Graph, enables companies to easily capture (mathematically encode) all engineers’ expertise around the globe on over 200 crude types and translate that into corrosion mitigation recommendations for all refineries around the globe.
The Computational Knowledge Graph is Maana’s invention; it is a new way to represent industrial knowledge mathematically. The unique innovation enables industrial companies to capture human expertise and data from across silos into digital knowledge to help employees make better and faster decisions. It is designed to be used by subject-matter experts to quickly develop AI-driven Knowledge Applications that accelerate digitizing operations. These Knowledge Applications are powered by models that provide recommendations into those operations. The Maana Computational Knowledge Graph™ shows the interdependencies between various entities that impact the operations and make it much easier to optimize hundreds of operations at a much faster speed.
In the case of crude oil, implementing Maana allowed the oil company to develop a Crude Flex Knowledge Application, which captures the expertise of subject-matter experts for use by all corrosion engineers. Using the various models stored in the Knowledge Application, the company’s engineers can make decisions that reduce maintenance costs and unplanned downtime.
In order to truly leverage digitalization, companies can simply look at the disruption Amazon has caused to the retail industry. Amazon on a daily basis is creating 6000 decision models per day on its customers and operations vs 200 created by an average retailer.
By establishing a CDO role, making AI an imperative, and selecting software that is capable of combining data with human knowledge, even the world’s largest organizations can create decision models at scale and become truly digital companies, improving operational efficiencies and adding billions of dollars to gross margins.