(US & Canada) VIDEO | AI Simulates Past Human Cognition to Predict the Future — AutoZone IT Management Director

(US and Canada) Himali Kumar, Director of IT Management at AutoZone, speaks with Asha Saxena, Founder and CEO of WLDA.tech, in a video interview, about the impact of Generative AI, what AI is, and the progression from narrow AI to AGI for future success.

At the outset, Kumar expresses her enthusiasm for data and engineering as an IT professional with over 20 years of experience. Her organization AutoZone is a US$16 billion auto parts retailer with 7,000 stores across the U.S., Mexico, and Brazil.

Emphasizing Generative AI, Kumar affirms that conversations about data are impossible now without mentioning ChatGPT. She highlights how it comes up in all executive decisions and boardrooms. She reveals even AutoZone is weighing its options for taking advantage of the generative conversation it can create.

Putting her POV on AI, Kumar refers to Gartner's analysis of using data – descriptive, diagnostic, predictive, and prescriptive. She believes these categories determine how to use data for problem-solving and making decisions that impact revenue.

Starting with descriptive, Kumar explains this is the lowest portion of the categories where the focus is on understanding reports. Moving into diagnostic, here data answers why changes occur, such as increases or decreases in sales in hours. This stage still belongs to Business Intelligence and Reporting, she adds.

Predictive and prescriptive go further, predicting what will happen in the future and suggesting solutions. Kumar affirms that the predictive and prescriptive categories fall in the realm of artificial intelligence.

In continuance, she asserts that, with artificial intelligence, humans are attempting to simulate human cognition from past occurrences to predict potential future scenarios. Should the same conditions exist, the system can prescribe what action should be taken to avoid or facilitate the predicted outcome. Kumar notes that in the past year, predictive and prescriptive aspects of artificial intelligence began to become more prominent.

Furthermore, she agrees to the progression from supervised to unsupervised, to reinforced learning in AI. For example, supervised learning often requires the input of a data scientist or analyst and the testing of a model for accuracy. In unsupervised learning, the machine has more control to determine data clusters, such as the estimated delivery time for an overnight package, reckons Kumar.

Ultimately, she sees a progression in artificial intelligence from narrow to general AI. She affirms that with the implementation of reinforced learning, the progress of AI is visible, and it is closer to the goal of general AI. She concludes by stating that understanding and utilizing the three kinds of AI will drive faster progress in the future.

CDO Magazine appreciates Himali Kumar for sharing her data insights with our global community.