GB710 Unit 6 Seminar Reflection Example

Sixty Percent of What? A Unit 6 Seminar Reflection on Tracing a Figure Everybody Repeats

[Student Name]

Purdue University Global

GB710: Contemporary Challenges in Business

Unit 6 Seminar Reflection

[Instructor Name]

[Date]

The candidate, the classmates and the intermediate sources are composite and illustrative; the original estimate and the sources cited are real.

Why the Figure Was Believed

When my classmate said sixty percent of an entry-level analyst's work could already be automated, my first reaction was to write it down. It matched what I had heard from partners at my own firm, which cut its graduate intake this year, and it seemed to confirm that the decision was sensible. I did not ask where it came from. Looking back, the figure was persuasive because it was precise, because it came from someone I respected, and because it supported a decision I had helped make. None of those is a reason to believe it.

Four Links in the Chain

Our instructor asked the group to trace the figure to its origin, and we spent about twenty minutes doing it. The classmate had seen it in a professional networking post by a technology consultant. The post linked to a trade magazine article about AI in finance. The article cited a white paper from a vendor of AI tools for accounting. The white paper, finally, cited a 2023 report from the McKinsey Global Institute on the economic potential of generative AI.

That report estimated that current generative AI, combined with other technologies, had the technical potential to automate work activities that absorb 60 to 70 percent of employees' time across the economy (McKinsey Global Institute, 2023). It was a sweeping, carefully hedged estimate about the global economy, not a statement about entry-level analysts.

Tasks, Potential and Adoption

The gap between the original estimate and the claim in the room had three parts. First, the report measured work activities, the individual tasks that make up jobs, not jobs themselves. Automating 60 percent of the tasks in a job does not remove 60 percent of the job if the remaining tasks still require a person. Second, the report measured technical potential, what the technology could do in principle, not what firms had actually adopted. Its own authors noted that adoption would take years and depend on costs, regulation and organizational change. Third, the report described workers across the economy, not entry-level analysts specifically. By the time the figure reached our seminar, each of these qualifications had been dropped.

The figure did not become false at any single step; it became false by losing a qualification at every step.

What Left the Running File

I keep a running file of claims and sources about AI and graduate hiring, which I am using to build my literature review. After the session, I went through it and found four figures I had recorded without checking their origin. I traced each one. Two came from consultancy estimates of technical potential, like the sixty percent figure. One came from a survey of employer intentions, reported as if it described actual hiring. Only one, from a payroll-based working paper, described what had actually happened to employment, and even it was a working paper not yet reviewed. I deleted the first three from the file's "evidence" section and moved them to a section labeled "claims in circulation," with their origins noted.

What I Learned About My Own Reasoning

The episode showed me that I was more careful with figures that challenged my firm's decision than with figures that supported it. Rousseau (2006) asks managers to look for the evidence most likely to overturn their view, and I had been doing the reverse without noticing. The sixty percent figure entered my file in seconds; a payroll study that complicated my firm's decision had taken me a week to accept.

A Sourcing Rule for the Remaining Units

I adopted a rule for the rest of the course: no figure enters my work unless I have read the original source, and every figure is reported with what it measured, tasks or jobs, potential or adoption, and for whom. Briner et al. (2009) describe evidence-based practice as asking what the best available evidence actually shows, and the rule is my way of applying that to my own file.

What Would Change My Mind

If a study using firm-level data showed that accounting firms adopting AI had in fact eliminated most first-year tasks, without shifting them to more senior staff, I would treat the high automation figures as closer to reality for my field. Until then, I will report them as estimates of potential. I also plan to share the traced chain with the partners at my firm, since the same figure has appeared in our own discussions of next year's intake.

References

Briner, R. B., Denyer, D., & Rousseau, D. M. (2009). Evidence-based management: Concept cleanup time? Academy of Management Perspectives, 23(4), 19-32. https://doi.org/10.5465/amp.23.4.19

McKinsey Global Institute. (2023). The economic potential of generative AI: The next productivity frontier. McKinsey & Company.

Rousseau, D. M. (2006). Is there such a thing as "evidence-based management"? Academy of Management Review, 31(2), 256-269. https://doi.org/10.5465/amr.2006.20208679