What generic AI gets wrong with this prompt
A 300-character box breaks most of generic AI’s habits at once. These are the expensive ones.
It spends a tenth of the box clearing its throat
“Upon completing my MBA at Booth, I intend to pursue…” — forty characters gone before any content arrives. Models pad openings because prose in their training data has openings. A short-answer box doesn’t; it starts at the substance or it wastes the question.
Ask your AI — “Count words in my answer before the first concrete noun. What could that space hold instead?”
It names the job and skips the why
“Return to consulting as an engagement manager” fills the box and says nothing distinguishing — titles are the most common completion in the model’s training, and the why-this-step is the part it can’t infer. What the role builds toward is the difference between a plan and a placeholder.
Ask your AI — “Does my answer say what this job builds or tests on the way to my longer-term goal? If not, which words would I trade for that clause?”
It hedges between two goals
“Consulting or a strategy role in tech” — offered ambiguity, the model preserves it. In 300 characters a hedge reads exactly like what it is: a decision not yet made. One goal, stated like you’ve already interviewed for it.
Ask your AI — “Does my answer commit to one path? If it contains an or, which side does the rest of my application actually support?”
It backs into the future from your past
“Leveraging my six years in healthcare consulting, I will…” — generic AI establishes credibility you already established elsewhere, because the material you feed it is all history. The resume proves the past. This box is the one place that only exists for what’s next.
Ask your AI — “How many characters of my answer describe what I’ve already done? What future detail did they displace?”