If you’ve been studying artificial intelligence and its growth, you’ll know that the industry is well past its nascent stage now. There is significant maturity in its growth, and companies from diverse backgrounds are realizing the impact of incorporating data and AI into their ecosystems.
In a bid to understand the dynamics of this data-centered growth, I teamed up with Hewlett Packard Enterprise (HPE) to do an analysis of their international survey on the present and the future of AI within the industrial sector. Through the survey HPE wanted to find answers to some of the most asked questions related to the field of AI:
What percentages of companies are working on AI?
How can AI transform industrial IoT for the better?
Will it prove to be a job killer for us humans?
The survey answers these and many other questions that shroud the mind of everyone who is currently a doubter.
AI has promised exceptional value addition to organizations that have incorporated it into the activities of their value chain. Not only has it enabled organizations to increase the efficiency of their operations, but it has also led to the creation of a better customer experience.
This progress has been enabled, among others, by the exceptional processing power and capabilities of systems that run AI, both at the ‘edge’, i.e. close to the industrial equipment, and in corporate data centers or clouds. In combination with years of experience working with analytics tools, and an increasing availability of AI frameworks, consulting and professional services, this has enabled organizations to use AI as a tool to add new features to their products and services, and center them on customer needs.
AI is deemed by many organizations as an opportunity to achieve certain aims and goals. Based on results from within the survey, we had a look at what some executives wanted to achieve through the use of AI in their business.
Interestingly, 57% of the respondents said that they wanted to increase their operational, supply chain, and maintenance efficiency through the use of AI. In addition, 45% wanted to improve customer experience and 41% were looking forward to enhancing the practicability of their offerings.
Industrial companies also wanted to use AI to increase employee productivity, and to create new business models, products, and services. This offers a hint that the end goal these organizations are seeking from using AI is rapidly changing. As time goes by, organizations are beginning to realize that AI has multi-faceted benefits, and should be used accordingly.
From the results of HPE’s survey, we learned that over 50% of companies said they are engaged with AI, with 11 percent having already implemented the technology in core functions or activities, 14 percent planning to do so within the next twelve months, and 36 percent evaluating the implementation. However, 39% of companies still had no plans to implement AI.
Additionally, respondents were asked about the obstacles to implementing AI. Nearly half of respondents consider the lack of data quantity and quality as a hindrance in the implementation of AI.
One example that springs to mind is a client HPE worked with, from the tobacco industry. They had five separate machines operating within the same area. However, one of them would fail every once in a while. It was extremely perplexing and confusing for the client, as they now had a machine that would fail numerous times, whereas all others in the same place operated well. The reason for the continuous failures in the machine, after endless research, was found to be present in the humidity levels in the factory. This caused issues with the metal used in that specific machine. Now, this was an issue that an organization would never be able to find out about. They couldn’t correlate humidity during the day with the type of metal. Fast-forwarding to now, these are the types of issues you can handle with the proper use of AI.
Moving onwards, 42% of all respondents believed that lack of AI skills and knowledge was a hindrance in the implementation of AI in their organization, while 34% also believed that lack of data governance and enterprise data architecture was an obstacle providing high quality data for AI models and stopping them from widespread implementation.
When asked about the departments they would like to implement AI solutions in, most executives pointed towards obvious answers, but in a rather interesting chronology: 38% mentioned they would use AI in research and development if they had the opportunity. Maintenance and operations departments were tied at 34% and 32%; services followed with 29% of people saying they would implement AI solutions in this department.
The interest in the use of AI in research and development has led to an interesting use case for AI.
Why are organizations currently bent on using AI in research and development, more than any other department? A couple of reasons for this could be:
While R&D received the most votes in the survey, respondents are deploying AI in broad range of departments and working together with partners in the ecosystem. An ecosystem stitches the model together and makes it work.
With AI expected to become a bigger and better part of the organizational ecosystem in the future, HPE’s survey also asked respondents which scenarios they were expecting from the technology by 2030. Almost 55% of respondents voted in favor of predictive manufacturing, which includes superior predictions of demand and other metrics.
Moreover, 45% voted for self-repairing and self-configuring machines, and 44% for mass customization. This is not only about making maintenance easy by repairing and configuring their mechanism themselves, but it’s a foundation for implementing key visions of Industry 4.0, for example order-driven production, and others.
When asked about AI phasing out human input by 2030, two-thirds of respondents mentioned that AI will not be a job killer. AI will enable us to work better, more efficient and do new things we have never been able to do. It also holds the potential for changing and disrupting business processes and models for the better. While results from the survey are positive, how it actually pans out is yet to be seen.