51�Թ�

During my 10 years as a college professor, my classes were nearly all graded primarily based on writing assignments. Though they had flashy titles—Contemporary Mexican Literature; Cervantes and the Invention of the Novel—their subject was always the same. The course material gave us topics for debate and analysis, but what I taught my students, or what I hoped to teach them, was how to improve as writers.

My training in rhetoric and composition studies, my own experience as a writer, and what I saw in my classes all pointed in the same direction. The way writing is often taught leads to a common misconception. Students think the essay is just the final draft, the version they turn in for a grade or submit with an application. To change this outlook, I designed my classes to emphasize the entire writing process. Essays were broken down into a series of scaffolded steps and deadlines. Each step built on the last one. Each had clear expectations and a separate due date. Each step in the process became obligatory, contributing points to their final grade. This helped students avoid procrastination and reduced the anxiety and pressure provoked by focusing on a complete, perfect final draft.

It just worked better. Better writing. Happier students. What I didn’t realize was that focusing on the process would become even more important after the advent of a new technology, capable of producing texts at almost unbelievable speed. This process-oriented approach will be particularly useful for independent educational consultants (IECs) now working with a generation of students learning to write alongside artificial intelligence (AI).

The Detection Trap

Despite this staged approach, some of my students still skipped the early process steps and, in recent years, turned in essays written by AI. I learned to recognize the signs, the telltale punctuation marks, negative affirmations, and endless sets of three, or worse, the completely invented quotations, or “hallucinations,” attributed to real course reading materials. My grades were mostly based on writing, so this felt like cheating. And like other professors, I tried out AI detection tools.

They didn’t work very well and the evidence around them suggests caution. A widely cited found that popular AI detectors flagged over 61 percent of essays written by non-native English speakers as AI-generated, while being nearly perfect for native speakers (Liang et al., 2023). The detectors score text based on “perplexity,” meaning that the simpler syntax common among multilingual writers gets confused for AI prose. Several major universities, including and , have since stopped using AI detection software.

Worrying about “detection” misses the point, anyway. The emphasis should be on process. Some college students have intuited this. A professor friend of mine told me the story of challenging a student about AI use, only for the student to prove they wrote the essay by sharing their Google Docs version history. It was a messy, ad hoc solution, but it was exactly the reframing of the problem endorsed by the MLA-CCCC Joint Task Force on Writing and AI in 2024, shifting the focus from detection to documentation.

The Opportunity

For IECs, this shift provides an opportunity to improve how we coach writing.

Focusing on the process leads to better results, and it also produces documentation. It helps students avoid the pitfalls and stresses of procrastination, and it also reduces anxiety for families who may be asking, “Will my child be able to prove that they wrote this essay?”

There is lots of evidence for a process-oriented approach. In 1981, Flower and Hayes’ cognitive process model showed that skilled writers move recursively among planning, translating, and reviewing rather than following a linear outline-draft-edit sequence. Nancy Sommers’ research showed that experienced writers revise by “re-seeing,” by rethinking arguments and restructuring ideas, while less experienced writers tend to focus on surface-level word substitutions.

But what does this process-oriented approach look like in practice?

A Five-Checkpoint Workflow for IEC Essay Coaching

The following framework adapts writing process research into a practical coaching sequence with five checkpoints.

1. Discovery

Before any drafting begins, the student does prewriting: freewriting, brainstorming lists, an idea map, exploratory notes. This is where the essay’s raw material originates. Encourage students to try multiple prewriting strategies, especially those that feel unfamiliar. The goal is quantity and honesty, not polish.

2. First Full Draft

The student writes a complete draft. The IEC’s role here is to resist the urge to intervene too early. Hattie and Timperley’s research on feedback effectiveness shows that more feedback is not automatically better. Premature comprehensive marking can overwhelm students and, paradoxically, reduce learning.

3. Revision Memo

Before receiving any feedback, the student writes a brief self-assessment memo: How well does the draft address the prompt? What still feels uncertain? What would they change if they had more time? This reflection externalizes the student’s thinking and gives the IEC a window into the student’s relationship with their own writing.

4. Focused Revision (Limited Priorities)

Respond to the draft with no more than three to four priorities. Resist the instinct to mark every issue. Research shows that selective, goal-directed feedback produces deeper revisions. Frame feedback as questions whenever possible, “What happens to your argument if you move this paragraph?” rather than “Move this paragraph.”

5. Final Polish and Reflection

The student completes a final editing pass for clarity, grammar, and voice. Then they write another memo, describing their writing experience. This closing step reinforces metacognition, the student’s awareness of their own writing process, which is the single most transferable outcome of the entire experience.

Protecting Voice

A final point, and maybe the most important one. Students who use AI to write their essays give up control over what most distinguishes them from their peers and fellow college applicants. They give up control of their voice. All AI writing sounds the same. The technology homogenizes, flattening linguistic diversity and cultural specificity, and what it produces is never funny or unique.

Admissions officers know this. That’s why authentic voice is the single most important criterion in college application essays. Here’s an easy technique to help students focus on voice. During revision, ask them to read their writing aloud and identify phrases and rhythms that sound most like their speaking voice. Is this how they really sound when they’re talking? Ask them to build a “keep list” of these phrases. The list will be useful and the exercise will also help them recognize their own authentic voice.

Some IECs may find themselves working with families that choose to try using AI tools. Here are four reasonable expectations for responsible use drawn from AI ethics frameworks (UNESCO, NIST):

• AI should coach, not generate content.
• AI interactions should be logged or documentable.
• Student writing should not be used to train AI models.
• AI should distinguish error correction from style flattening. It should never push students toward a single “correct” style or tone.

Institutions are already focusing on these priorities and IECs who adopt them now will be ahead of the curve.

By Jeff Peer, PhD, Founder and Coach, Profsy Writing Studio