David Menninger's Analyst Perspectives

In the technology industry, 2023 will be remembered as the year of generative artificial intelligence. Yes, the world was made aware of GenAI when ChatGPT was publicly launched in November of 2022, but few knew the impact it would have at that point in time. Since then, GenAI has taken the world by storm, with vendors applying the technology to make it easier to ask questions about data, write code (including SQL), prepare data for analyses, document data pipelines and use software products more effectively. More than three-quarters of enterprises are using or plan to evaluate GenAI across these different types of usage. As I’ve written previously, natural language analytics may finally become a reality. 

Unfortunately, all the focus on GenAI may divert attention away from “traditional AI” or predictive analytics. One could say that GenAI is sucking all the air out of the room. I anticipate that in 2024, GenAI will continue to enhance many aspects of business processes, but in the coming year I think we will also see a backlash about some of the shortcomings of GenAI relative to traditional, predictive AI.  

Because GenAI has been so successful at many of the tasks associated with data and analytics, it is perceived as a panacea. However, I do not anticipate that GenAI will dominate the use cases typically associated with predictive AI such as predictive medicine, fraud detection, predictive maintenance, intrusion detection and other similarly complex, highly valuable use cases. ISG’s Banking study confirms this split, and ISG’s AI Buyer Behavior study further reinforces this notion, showing that enterprises expect that one-half (51%) of AI spend in 2024 will be on predictive AI, and one-half (49%) will be on GenAI. 

AI was a hot topic before ChatGPT. Our research showed that three-quarters of organizations planned to increase investment in AI and machine learning, and none planned to decrease investment–and that was before the GenAI phenomenon. Unfortunately, as I’ve written previously, enterprises don’t have the skills to support the application of traditional AI/ML to operations. Only one-quarter of organizations indicated having the skills needed, and only one-third (31%) cited having adequate AI/ML technology. This problem is likely to continue. We assert that by 2026, more than one-half of enterprises will realize AI competencies and skills are insufficient and will require new investments to avoid being at a competitive disadvantage. 

What does this mean for enterprises? While you need to invest in GenAI and the benefits it brings to all types of business processes, you also need to continue investing in good old-fashioned traditional AI/ML and supporting data technologies. But it’s not just about technology; enterprises also need to address the shortage of skills to use traditional AI successfully. Eventually, GenAI will make it easier to develop predictive AI models, just as it makes it easier to develop code. But for the near term, the highest value AI/ML use cases will still require a human in the loop, armed with specialized skills and technology. 

Regards,

David Menninger