
Austin Tech Connect: The Podcast For The Austin Technology Ecosystem, Business Leaders, and Tech Entrepreneurs! Behind the Data on AI at Work, with Dr. Nick Hallman
In this episode of Austin Tech Connect, Thom Singer sits down with Dr. Nick Hallman, professor at The University of Texas at Austin, to talk about one of the biggest business questions of the moment, how companies are actually using AI, and whether they are measuring success the right way. Drawing on research conducted with KPMG, Dr. Hallman shares what his team learned from studying real workplace interactions with large language models over time.
What makes this conversation especially interesting is that the study did not just look at whether employees were using AI. It looked deeper. Dr. Hallman and his colleagues were able to examine prompts, responses, and patterns of use across months of professional activity. That gave them a far richer picture of what productive AI adoption really looks like, and what many organizations may be missing when they focus only on usage volume.
One of the biggest surprises? Most AI use was not especially advanced. Dr. Hallman explains that roughly 90 percent of the activity they observed was centered on writing help, things like cleaning up emails or improving wording. Useful, yes. Transformational, not really. The more sophisticated uses involved clearly defined tasks such as analysis, coding, and creating tangible work product. Those higher value outcomes tended to come when users were specific about what they wanted and what a successful output should look like.
Another unexpected finding was that senior people often used AI more effectively than junior employees. Dr. Hallman suggests that may be because strong AI use mirrors strong delegation. Leaders who know how to clearly assign work to people are often better at directing an LLM. That insight challenges the assumption that younger workers will automatically be the most advanced AI users just because they are more comfortable with technology.
The conversation also explores how companies should think about training and evaluation. If raw usage is not the best metric, what is? Dr. Hallman points to more meaningful signals, including whether people iterate with the model, refine their requests, and move beyond one-shot prompts. He also stresses that the best way to improve is through practice. The more people experiment, at work and at home, the more they begin to understand what AI can do well, where it falls short, and how to ask better questions.
This episode is a smart, grounded look at AI in the real world. It is not hype, and it is not fear-based. It is a practical conversation about what happens when organizations move past buzzwords and start paying attention to how people actually work with these tools every day.
About the Guest Dr. Nick Hallman is a professor in the McCombs School of Business at The University of Texas at Austin, where he teaches data analytics and Python to accounting students and conducts economics-based research related to the auditing profession. In this conversation, he shares insights from a research collaboration with KPMG focused on how employees are using AI in professional settings.
Key themes from this episode
- AI usage numbers do not tell the whole story.
- Most workplace AI use is still basic writing support.
- The best AI results come from specificity and iteration.
- Senior leaders may be better AI users because they are better delegators.
- The fastest way to become more effective with AI is simple, use it more often.
Sponsor: Austin Tech Connect is supported by Calavista Software, software development without the drama. Trusted by startups and Fortune 100 companies alike.
