How to Use AI Effectively in Writing Letters of Recommendation

About 10 years ago, I had a student who came to me and wanted to apply early decision to a highly selective college. By looking at her midterm grades, I knew that getting accepted would be a challenge because she wasn’t as strong in math as in her other subjects. I was aware that she was putting in work behind the scenes to be successful in the two extremely rigorous math courses she was taking. She was working with a peer tutor, attending study groups, and meeting with the teachers after school. Yet, I only had this information because I knew her personally. That was not something an admissions committee would see or understand by looking at her grades.

I strongly suggested that she get recommendations from her two math teachers to provide additional context about rigor of the class, the class average, and how she had been working hard to improve. She was accepted for early decision. Later, in a conversation with a college representative, he told me that the additional information about math courses tipped the scale. As counselors and teachers, we are storytellers who provide color, context, and depth beyond the grades and testing.

As technology has evolved, I’ve begun to incorporate artificial intelligence (AI) into my workflow. However, when it comes to recommendations, that is a delicate process. Teachers and counselors know the key personal details about students that can make a real difference in the application process. While I use AI tools in some parts of my process, there are some aspects that remain (and for me will always be) human-driven. In this post, I’ll share some examples and resources that can help make writing letters of recommendation easier—with and without technology.

Using Surveys and 1-on-1 Meetings to Collect Information

When I started advising students, I asked them to complete brag sheets or autobiographies. These surveys allowed students to share their interests, strengths, activities, family ties, community service, and academic successes. Consider trying this method to get unique information based on students’ personal views of themselves.

While surveys are very useful, the human element of face-to-face interaction is unmatched. During my second year as a counselor, I inherited 30 seniors from the coordinator for school counseling. I had a short window to become acquainted with the students before writing their recommendations, so I interviewed each of them. These meetings were productive and insightful—I was able to make connections with them and gain direct insight about their stories, which yielded important insights that could be shared in a letter of recommendation.

Essential Elements of a Recommendation

According to the National Association for College Admission Counseling, both the teacher and counselor recommendation are highly valued in the admissions process. While there are some similarities in how a teacher’s or counselor’s letter of recommendation is structured, the main difference between them is the context they provide.

Teachers observe how students interact with each other and how they’ve changed during the year. This context is important to the admissions committee because they are looking for information about the student and who they are in the school and community. After spending many months with students, teachers can provide specific and valuable information about students’ strengths or challenges, how they contribute to the classroom environment, and their abilities as a leader or teammate. If teachers have served in multiple roles as a coach or adviser, they can also provide additional nuanced details about the student.

Ethan Sawyer, the College Essay Guy, provides resources for teachers to use as they write letters of recommendation.

The counselor recommendation is holistic and provides context about who the student is in the school (involvement in the school and local community, including clubs, sports, music, the arts, employment, internships, and leadership). If they don’t have time to get involved at school because they help out at home or work, make sure to share that information in your letter. The College Board also offers useful tips for counselors in the process of writing letters of recommendation.

It’s important to note that personalization for letters of recommendation is always best when possible. Additionally, make sure that you don’t repeat the same information that’s easily found on students’ résumés or transcripts.

Formatting: Bullets or Paragraphs

For many years, I wrote college recommendation letters that were one to one-and-a-half pages. I included stories and narratives that highlighted students’ multidimensionality, such as notable details about being first generation college students, overcoming adversity, working while in school, babysitting, playing an instrument, achieving success as an athlete, or expressing artistic and creative talent.

Although a formal narrative has been the standard for recommendation letters, it’s not a requirement. A few years ago, I spoke to a number of admissions professionals at different competitive schools who all agreed that bullet recommendations were fine if the specific context was provided. In The Journal of College Admission, Trevor Rusert and his colleagues provide a useful approach that educators can try.

Using AI Tools

At a conference last year, during my session titled “Let’s Compare Notes on Technologies and Processes,” I asked a room full of counselors and educators about their experience using AI to assist them in writing recommendations. The room was quiet. I could see that people were hesitant to openly discuss their thoughts about using AI.

AI can be very helpful for counselors and teachers who have a large number of letters to write. My caseload included 60–70 seniors, five to 10 additional letters from students I mentored, and 160–180 students in ninth through 11th grade. With a truly considerable amount of letters to write, a mentor recommended Perplexity.ai to me. During my last two years as a counselor, I used it to check grammar in some sentences and to brainstorm different ways to present an idea. I always made sure to craft the personalized content for each student myself.

AI can generate ample suggestions, but caution is important when we’re using it to help us refine letters of recommendation. Protect students’ privacy by making sure not to paste their names (or personal information about their family) into the tool you’re using. Also, avoid using a template for every student. This was a common practice for a few educators (not the majority) when I was an admissions director, but students inevitably ended up receiving the same recommendation letter. AI can be effective for organizing your thoughts before writing, so some might benefit from using it to assist them with creating a general outline.

Technology is a Tool, Not a Witness

When I started writing recommendations many years ago, I used a typewriter and mailed all letters to colleges and employers. Today things are very different; communication in this process is digital. Can AI write a letter for you? Yes, it can.

However, AI can’t replicate the experience of watching a student change and grow over the years. Technology can’t replicate what we witness. Technology and AI tools are constantly changing and widely used, so continued professional development and research is essential so that educators can consider best practices when writing letters of recommendation.