Write an Honest Letter of Recommendation from Your Own Notes

You get a ready-to-send recommendation letter where every strength is backed by an example from your notes, plus a claim check table and a list of weak spots.

Prompt

Prompt
You are an experienced manager who writes specific, honest recommendation letters that hiring committees trust.

Write a letter of recommendation from the notes I will paste after this prompt. First ask me, in one numbered message: 1. What is the letter for, and who will read it? 2. How do I know this person, and for how long? 3. Is there a length limit or are there required points? Wait for my answers, then write.

Rules:
- Use only facts and numbers from my notes. Never invent achievements, quotes, metrics or rankings.
- Tie every strength to one concrete example from my notes in the same paragraph.
- Pick the 3 qualities that best match what the reader wants. Drop the rest.
- Do not use generic praise (hardworking, passionate, team player) unless an example follows it right away.
- Write in first person, in a plain professional tone, under 400 words.

Output format:
1. The letter: greeting, an opening that states our relationship and a clear recommendation, 2-3 body paragraphs, and a closing that offers to be contacted.
2. Claim check table: Claim | Example from notes.
3. Weak spots: claims with thin or missing evidence, each with one question I could answer to strengthen it.

If the notes cannot support a claim, leave it out and list it under Weak spots instead of filling the gap.

Example output

Example for: a team lead recommending a former junior data analyst for a Master's in Data Science, 2 years working together.

Letter

Dear Admissions Committee,

I managed Priya Nair for two years on the analytics team at Northfield Logistics, where she worked as a junior data analyst. I recommend her for your Master's in Data Science without reservation.

Priya teaches herself what a problem needs. When our weekly delivery report took six hours to build by hand, she learned Python on her own time and automated it within a month. It now runs in under 10 minutes, and the team has used her script every week since.

She also questions numbers that look too clean. In her first year she noticed that our on-time delivery rate excluded rerouted shipments. She raised it with me, documented the gap, and helped rebuild the metric. The corrected figure dropped from 94% to 87%, and that honest number led us to renegotiate two carrier contracts.

Finally, she explains technical work to people who do not want to read it. Priya ran four lunch sessions teaching warehouse supervisors how to read our new dashboard. Requests to our team for basic report pulls fell noticeably afterward because supervisors could find the answers themselves.

Priya is ready for graduate-level work, and I expect her to make the most of it. I am happy to answer any questions at the contact details below.

Sincerely,
Daniel Okafor
Analytics Team Lead, Northfield Logistics

Claim check

ClaimExample from notes
Teaches herself what a problem needsLearned Python and automated a 6-hour weekly report down to 10 minutes
Questions numbers that look too cleanFound rerouted shipments missing from the on-time metric (94% to 87%)
Explains technical work clearlyRan 4 dashboard sessions for warehouse supervisors

Weak spots

  • Fewer report requests: your notes give no number. Roughly how many requests per week came in before and after the sessions?
  • Academic readiness: nothing in your notes covers statistics or math. Did she do any modeling, forecasting or statistical testing you saw firsthand?
  • Leadership: you mentioned she "stepped up" but gave no example, so I left it out. Was there a time she led a project or mentored someone?
Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • Requiring one concrete example per strength turns vague praise into evidence that admissions and hiring readers actually trust.
  • The claim check table lets you verify in seconds that nothing was invented beyond your notes.
  • The Weak spots list with follow-up questions shows exactly what to add instead of letting the model pad thin areas.

Use this when someone asks you for a reference letter for a job, grad program or scholarship. Paste your rough notes and examples into ChatGPT, Claude or Gemini right after the prompt, then answer its three questions.

Common edits

  • Change the word limit to 250 for an online reference form with a character cap.
  • Add the job posting or program description so the 3 chosen qualities match it more closely.

Prompt details

By
Prompts-GPT editors
Published
How prompts are written and checked

Did this prompt work?

Keep exploring

More in Writing