(Maybe) Why Experienced Managers Excel at AI Prompting
A few days ago, Ethan Mollick, a professor at Wharton (Penn University) studying AI, innovation and startups, posted the following on X:
"I keep hearing from executives that they expect that a new generation of "AI natives" will show them how to use AI. I think this is a mistake: 1) Our research shows younger people do not really get AI or how to integrate into work 2) Experienced managers are often good prompters" - link
This got me thinking about how hard good managers work at the wordsmithing required to achieve clarity in the workplace. Particularly prevalent post-pandemic is written communication due to the shift to remote work, part-time or otherwise. Crafting precise instructions, objectives, constraints, or feedback is a skill experienced managers often get to exercise under these conditions as success depends on it. Many of us have learned (or certainly will) the hard way that George Bernard Shaw was 100% correct when he said: "The single biggest problem in communication is the illusion that it has taken place."
Younger employees are often expected to be "AI natives" due to their familiarity with technology and social media platforms. However, while they may know how to use these tools, their understanding of strategically integrating AI into complex work environments is less developed. In contrast, experienced managers have had years to hone their communication skills through trial and error. It is not surprising to me that these written communication skills translate well to AI prompting as I am reminded of a silly poem in one of my undergraduate computer science textbooks that has proven true repeatedly in my software engineering career:
"I really hate this damned machine, I wish that they would sell it,
it never does quite what I want, but only what I tell it!"
This rings especially true with Large Language Models (LLMs), where even slight differences in phrasing can drastically change outcomes. Since these systems are fundamentally probabilistic word prediction engines (a simplification but valid for this discussion), at a basic level, LLMs receive input, tokenise it (breaking it into pieces), and use those pieces to predict the most likely next token to produce the output. In this context, it makes sense then that seasoned managers, who have mastered clear articulation through experience, have an edge in "prompt engineering". However, this advantage might not last as LLMs and AI systems improve in understanding vague or ambiguous inputs.
Of course, integrating AI into the workplace is not solely the domain of any single age group. While experienced managers bring strategic insight and communication expertise, younger employees contribute enthusiasm, adaptability and out-of-the-box thinking. The true disruption of AI in the workplace will only come once we leverage it not merely as an enhancement to our current ways of working but as a means to unlock novel ways of working previously unimagined.
I bet it won't be the experienced managers who figure out the cheat codes.
Additional Reading & References
Google | Overview of Prompting Strategies
Atlassian | The Ultimate Guide to Writing Effective AI Prompts
OpenAI | The Art of AI Prompt Crafting: A Comprehensive Guide for Enthusiasts