Since the 1980s, I have been engaged in experiments involving artificial intelligence systems. During that era, the use of computer systems to attain groundbreaking capabilities was viewed as an emerging trend in the field of AI.
Today, with the emergence of generative AI (genAI), the landscape has remained relatively stable. However, organizations must have a clear grasp of the essence of AI to avoid the errors that were prevalent in the early stages of AI methodologies.
Reflecting on 1988
It would be unfair to compare the capabilities of traditional AI technologies from the 1980s, such as Lisp and M1, with modern machine learning and genAI technologies. The costs associated with AI methodologies back then were prohibitive, and their effectiveness was significantly lower.
Numerous mistakes contributed to the decline of AI, while businesses opted for simpler solutions. One prominent error was the inappropriate use of AI in situations where its utility was limited.
Even in my initial experiences, I acknowledged that tasks like sales data entry were not suitable for AI. Despite this realization, I was authorized to implement such systems, fully aware that it was like using a sledgehammer to solve a minor problem—at a substantial expense.
This misalignment between the capabilities of AI and practical business applications largely led to the waning interest in AI within corporate circles. Nevertheless, in the subsequent decades, AI has made a comeback in the forms of conceptual AI, deep learning, and machine learning.
Despite significant technological progress and cost reductions, I still witness perplexing mistakes being made. The deployment of costly genAI systems developed by high-priced experts often fails to deliver the expected returns, potentially resulting in disillusionment in the near future if corrective actions are not taken.
These self-inflicted setbacks can be entirely avoided if companies are cautious in utilizing this technology. What are the primary genAI corporate initiatives? Which applications are considered suitable or unsuitable? How should organizations navigate the selection process? And how can the mistakes of the past be prevented?
By examining the proficient functionalities of genAI and identifying the use cases that highlight these capabilities, one can determine where to refrain from its implementation. This decision-making process does not need to be overly complex.
For the purposes of this discussion, let’s concentrate on the top three applications. While this list is not exhaustive, it suffices for the scope of this blog post.
Mastery in Natural Language Generation
The primary application is NLG, which refers to natural language generation. If you have ever tried to pass off a document, letter, email, or any written content generated by ChatGPT, you are already familiar with this technology.
Businesses can utilize NLG to create significant value, especially in improving customer interactions through personalized written or chatbot-driven communications.
While this automation may lead to job displacement, the ability to achieve more with fewer staff can be beneficial for enterprises. Enhanced customer experiences and expedited issue resolution can be achieved through NLG.
Imagine contacting a technical support helpline equipped with interactive voice response (IVR) systems; the effectiveness of the interaction largely depends on the knowledge and communication skills of the representative. What if an entity had the collective knowledge and expertise of 10,000 experts, enabling them to address customer queries swiftly and effectively? This highlights the potential for NLG to provide superior value and enhanced customer experiences at a reduced cost, if implemented wisely.
Personalized Advisory Systems
Advisory systems involve genAI-enabled platforms tailoring recommendations in e-commerce, streaming, and content dissemination platforms. While this concept is not new—I have been refining these systems before the advent of genAI—the current technology allows for a higher level of effectiveness. These systems offer unmatched return on investment for retail businesses.
Have you ever wondered how an online retailer can suggest products to you even before you disclose any preferences? By understanding the user’s demographic details, preferences, and occupation, these systems—previous versions of which could boost sales by 20% to 40%—can recommend products and services customized to the individual’s likely needs.
With the evolution of genAI, businesses can engage customers through meticulously crafted, dynamically generated connections, achieving an unprecedented level of effectiveness. From personalized typography and subliminal messaging to color schemes and pricing strategies, these systems excel at stimulating consumer interest and driving sales. Prepare for a marketing blitz aimed at bolstering the bottom line, albeit raising ethical concerns in the process.
Enhanced Anomaly Monitoring
Anomaly monitoring involves identifying irregular data patterns or outliers for applications like fraud detection or system surveillance. In this domain, genAI can help in recognizing data trends, elucidating their implications, and refining processes to optimize business outcomes.
This surpasses the traditional anomaly detection phase of genAI, which uses historical data patterns to flag potential fraud or predict system disruptions. It resembles a scenario akin to “pre-crime” from the movie “Minority Report,” where your future loan application could be rejected based on predictive analytics. This application underscores the societal implications that require careful consideration.
While there are numerous other compelling applications for genAI, the key lies in companies recognizing these opportunities and avoiding scenarios where genAI could escalate costs and risks without commensurate benefits. Prompt action is crucial to prevent self-inflicted wounds that could potentially jeopardize the future of businesses.
