Hello everyone, My name is Kenshin Fujiwara and I am the CEO and founder of HACARUS Inc.
Through this series of blogs, I will discuss a wide range of topics of AI, from the history of AI to practical tips for successful application of AI projects. I hope my blog posts will help you gain a better understanding of AI and solve your business issues.
In the last blog, we studied common barriers that interfere with AI application in existing systems in the business field. Today’s blog is the second part of it. I will continue to discuss two other barriers in AI application: the KDD barrier and the cost-effectiveness barrier.
The KDD Barrier
KDD (Knowledge Discovery in Databases) refers to the importance of reconciling intuition, experience, and grit, which are all valued in the field. One example is the world of Japanese craftsmanship. Many of these craftsmen have specialized skills and we refer to this as “on-site power.”
When introducing AI into these environments, cameras and IoT sensors are installed on the equipment and tools used. The collected information is then converted into data, a process that is difficult for humans to do. Using this method, the AI can use the data to optimize operations, simulate when machinery will break down, and when parts need to be replaced. However, even in these environments, the barriers of KDD are still there.
Cooperation between AI and Craftsmanship
When working with production managers, it is easy for them to say, “We don’t need to introduce confusing and complex AI into our work because our workers’ experience and intuition is superior.” From their point of view, it is inevitable that there will be some people who reject AI because it is “unknown.” I have also seen resistance in the field when introducing AI significantly changes how their current workflow function. Even after explaining it, some workers don’t like the idea of change. They feel like it will somehow increase their workload, workers lose their jobs, and increase their operating expenses.
The first step to overcome such hurdles is to show them how useful AI is through successful use cases within the company. These use cases can be small or simple, as long as they are successful. This method works well for cases where AI adoption is low. Showing how AI can reduce work hours or increase productivity has been very effective in changing a company’s opinion on AI implementation. After one successful project, news of the benefits will propagate throughout the company. Other departments can be swayed this way, and field workers can also begin to accept the use of AI.
The Cost-Effectiveness Barrier
Even if an AI system is implemented and operates effectively, it will be meaningless if the benefits to the business are less than the costs of operating it. As mentioned above, operating an AI system requires a lot of data which also costs money.
The first step is establishing a mechanism for collecting and analyzing the data. For example, suppose data is extracted from an existing internal system, such as Enterprise resource planning (ERP), and is processed using AI. In that case, it is necessary to build a new system for the analysis. This could involve processing the data sources or storing them as a database.
In today’s business world, cloud services help minimize the initial cost of implementing AI and shorten the building time. However, these costs are still costly, and the systems still incur daily operating costs. They also require human resources to manage and operate them.
While AI is a powerful tool, it isn’t always the most effective solution. There are cases where the isolated AI system is cost-effective, but the entire business process leads to lower productivity. Looking at an example, suppose that AI is introduced to one part of the business process, and the productivity of that process is increased.
While this might seem positive, because the processes before and after the AI-assisted process were not digitized, extra administrative work is required to handle the inputs and outputs of the AI. This can result in lower overall productivity for the system as a whole. Therefore, when introducing AI, it is necessary to review the entire business process for end-to-end overall optimization.
Difficulties in Calculating the Cost-Effectiveness
At the same time, it is difficult to calculate the cost-effectiveness of some projects. As mentioned above, AI can be used to optimize operations and replace human tasks with AI. The main objective is to improve efficiency and reduce costs, both of which can be measured.
In other cases, where AI is used to create new business or add value to existing practices, it may be difficult to predict the business benefits that will be gained. Even if this is the case, it is still necessary to calculate the cost-effectiveness, using the information available to calculate the sales forecast.
In terms of cost-effectiveness, a combination of AI implementation and public cloud services may also help reduce costs. However, since this type of AI is ready-made, it is important to determine whether it is the right fit for reaching performance goals and value targets.
In the next blog, I will continue to discuss the last barrier, the human resource barrier, which is used to support everything and introduce sparse modeling as a new solution to solving existing deep learning problems. For your updated knowledge and insight about AI technology, subscribe to our newsletter or visit HACARUS website https://hacarus.com.
