Columbia Business School MBA Admissions 2026-27 — Deadlines & AI MBA Essay Review

CBS essay examples: what generic AI gets wrong with the Essay 1-Dream Job

Columbia wrote the word dream into a goals prompt on purpose. Generic AI, trained on cautious goal statements, edits it right back out.

It answers “dream job” with a job title

“My long-term goal is to become a general manager in consumer tech.” That’s a job title. The prompt asked for the dream. Columbia invited imagination, and a dream job you can’t picture yourself inside, the problems that would sit on your desk, the people affected by your calls, reads as one you couldn’t speak about with passion in an interview.

Ask your AI — “Using only my draft, what does one working day in my dream job actually look like?”

It replays the resume the prompt just set aside

Columbia opens by saying it already has a clear sense of your path to date. Generic AI spends valuable word count restating work experience if you provided resume points. Without real expertise in what admissions officers look for, it struggles guiding you to extract tangible reasons why your 3-5 year goal will position you for success in your dream job.

Ask your AI — “What are tangible growth experiences that connect my goals? [drop in your goals, the more detailed the better]”

It misses the why behind your dream

Generic AI reads “dream job” as the biggest possible title, CEO, partner, managing director, because models equate ambition with titles. It lacks context to understand that the why behind your passion for your dream is a vital piece of the essay. It will more likely push for details around the role rather than ask to inject a reason that shows who you are and why you value this dream.

Ask your AI — “Is the passion for my dream job clear?”

CBS essay examples: what generic AI gets wrong with the Essay 3-Optimal MBA Experience

Columbia is asking you to design your two years from the school’s actual offerings. A generic essay writes about collaboration in general, and names nothing at Columbia you could actually impact.

It treats “special place” as a compliment to return

A generic draft opens with something like “Columbia’s collaborative culture and tight-knit community are exactly the environment where I would thrive.” The praise repeats what the committee already knows about its own program, and the sentence’s vivid words, “tight-knit”, “thrive”, are spent describing Columbia rather than you. That is the essay’s balance tipped the wrong way toward a description admissions readers do not need. Models write this way for the same reason a phone’s autocomplete finishes “thank you for your” with “time”. Given the prompt’s warm account of Columbia, the most likely next words are more warm account of Columbia.

Ask your AI — “Which of your sentences say something about Columbia that is not backed by something specific about you?”

It answers “Please be specific” with categories

“At CBS I would join entrepreneurship clubs and take negotiation classes.” A sentence like that could be sent to any school in the country. Even when a model names a real club, it tends to make the club the subject of the sentence and you the bystander in it. The deeper problem is the workflow. Asked for an essay, a model hands you a finished one, without guiding you to reflect on what you can uniquely contribute. An AI tool helps at this step when it asks you questions instead of writing for you.

Ask yourself — “If asked to speak to these contributions during an interview, would I sound full of passion? Do I have details to share about what I would do within that club, activity or class?”

It pads the page with wins the theme never asked for

When the model runs low on real material it reaches for your resume. A promotion or a product launch appears mid-essay, impressive on its own and unconnected to the optimal experience you described. Strong storytelling is achieved in this essay when you have a central theme like “the organizer in my community” or “the collaborator who gets quiet teammates heard”, and every example “shows” this is an authentic characteristic. A generic draft fills the space with accomplishments the committee can already read on your resume page.

Ask your AI — “What does this accomplishment add to the experience you say you want to build?”

CBS essay examples: what generic AI gets wrong with the Essay 2-Collaboration & Inclusivity

Columbia offers three kinds of team contribution and asks for a specific example of one. A generic draft claims all three at once, then recounts the outcome to the project or company instead of your part in changing the team.

It claims all three changes at once

“I made our team more collaborative and inclusive, and built a real sense of belonging.” The prompt lists three options and a generic draft takes the whole menu, because covering everything the prompt mentions is the safest answer a model can produce. It won’t warn you that an essay this short proving three changes proves none of them, and one proven impact is what the committee is reading for.

Ask your AI — “Which single change does my example actually prove, collaborative, inclusive, or community?”

It rewrites your team change as a leadership story

“As project lead, I directed the team toward a more open culture.” The prompt never asks whether you were in charge of a team, and Columbia has other essays where being in charge belongs. The model reaches for authority because most team stories in its training data are told by the person running the room. What this question wants is influence between people you touch (peers, mentees, teammates).

Ask your AI — “Would my example still work if I removed every mention of my title or role?”

It spends your words on the project instead of you

“The client deadline was tight, the workstreams were siloed, and morale was low.” Scene-setting like that eats the word count without focusing on your actions. The model defaults to context because context is the easiest part of any draft to write; it can describe a project without knowing you at all. Columbia asked for a specific example of something you did, and every sentence about the situation or your colleagues is a sentence that is not about your moves.

Ask your AI — “How many of my sentences have me doing something, rather than describing the situation or other people?”

It shows the actions but never why you cared

“I set up weekly cross-team lunches and rotated who ran our meetings.” Reasonable, and on their own they read as tactics anyone could copy. The model cannot supply the reason you noticed the problem and acted, because that reason exists only in your memory, so it substitutes a plausible business case. The reason is what makes these moves yours, and it is what turns office logistics into a sense of community the committee can believe.

Ask your AI — “Where does my draft say what made me care enough to act, before the actions start?”