Artificial Intelligence Could Soon Make Management Decisions

At the “Deeply Responsible Leadership” conference in London on March 15 hosted by the SC Johnson College of Business, the University of Oxford, and the Cornell Club of the United Kingdom, Vishal Gaur began his talk by explaining how large language models work by converting natural language into mathematical numbers, and back again.

“Generative AI is amazing because this software tool is actually created with 175 billion parameters. Can we even imagine that? We usually do regressions with far fewer parameters [in statistics] using much less data, and GenAI has been trained on 570 billion words.”

Gaur is the Anne and Elmer Lindseth Dean of the Samuel Curtis Johnson Graduate School of Management, and he is also Cornell’s policy advocacy group representative for the AI Alliance, a global network of universities, corporations, and scientific organizations creating standards for artificial intelligence.

Gaur said that within a business context, AI has led to the creation of chatbots that can converse with customers and helped us with organizational tasks and financial transactions, all within the last 18 months or so. The opportunities span healthcare, education, human resources, and beyond.

“On the one hand, we can automate tasks, we can enhance the productivity of our employees, and we can increase the speed at which they do things so we can reach a larger audience faster. So this is a great enabler for our organizations,” he said.

However, there are other kinds of AI beyond GenAI, and the potential for AI to make business decisions in management, consulting, and real estate, for example, is even more interesting, he said. Instead of uploading your own information to ChatGPT, for example, AI software could extract data directly from corporate databases.

“Think about that – a chatbot with access to your databases. How do we want to manage that?” he asked. “The future is not a future we can put away. It’s already a given.”

Gaur said that automations currently performed by straightforward rule-based algorithms will soon be done by AI, such as order fulfillment and robotic automations in warehouses at large online retailers that operate on a scale economically infeasible for human labor.

Imagine an employee’s headset is not working properly, Gaur said. The employee creates a service ticket. The AI software automatically connects to the company’s database, finds and orders a replacement, and tells the employee that it will be delivered tomorrow by 7 am. Then, it analyzes other customer service data, alerts a manager if complaints are significant, and suggests next steps. This is just the beginning.

“To do all this, we need to invest in different types of skills,” Gaur said. Leaders need to be ready to create the digital infrastructure for decision making, create custom algorithms within our organizations and think carefully about how our databases are connected with them, the security and privacy boundaries needed, and what is open source.

AI in Hospitality

As a PhD student in the 1990s, Chris Anderson, now professor of services management at the Cornell Peter and Stephanie Nolan School of Hotel Administration, programmed his first neural network to try to understand the interplay between prices and timing across different flights to help a particular airline maximize revenue.

“Neural networks are the foundation of our interactions with artificial intelligence,” Anderson said during his talk. Machine learning and artificial intelligence look for patterns, and for many decades, patterns were found in numeric data.

“Now, the language of the patterns has changed,” he said. Instead of simply numbers, the language is “everything,” including text, video, audio, molecular structure, and more.

In hospitality as in many other fields, AI can help automate tasks and enable more efficiency. So business leaders need to think about what data is meaningful, how to get it, and then use it to automate processes or make decisions. Specifically, AI could be used to customize content for guests arriving at a resort or, as Anderson has taught his students, to analyze brand perception across social media.

“But for me, it’s not about the kind of things we can do. For me, it’s about who can do it,” Anderson said. “It’s not just about a recent graduate coming to work and being a little bit more data savvy than you are. It’s about having these tools at everyone’s disposal.”

This article is part of a series:
Part 1: AI Will Revolutionize Business, Business Education, and Democracy
Part 2: Artificial Intelligence Could Soon Make Management Decisions
Part 3: AI Opens New Frontiers in Finance