Career

How to Write a Recommendation Letter with AI

Draft a specific, credible recommendation letter that genuinely strengthens a colleague's or student's application.

Generic recommendation letters do more harm than good. The most effective ones cite specific projects, include quantified outcomes, and make a direct argument for why the candidate is exceptional — not merely competent. AI can help you structure and draft a compelling letter that feels personal and specific even when you are short on time.

Why generic letters actively hurt applications

Admissions committees and hiring managers read hundreds of recommendation letters per cycle. They are expert at identifying the ones that are genuinely written by someone who knows the candidate versus the ones that are template-filled with the candidate's information. A letter that says 'Jane is a hardworking and dedicated student who contributed meaningfully to class discussions' tells an admissions reader nothing distinguishing about Jane — it describes half the applicant pool. Worse, weak recommendation letters signal to reviewers that the writer could not find specific evidence of distinction, which raises questions about whether the candidate actually has any. A letter's job is to make a specific, evidence-based case that this person is exceptional in a way that the resume and transcript cannot fully capture.

The structure that makes letters compelling

The most effective recommendation letters follow a three-part structure that mirrors a legal argument. First, establish the writer's credibility and relationship: how long, in what capacity, and what this gave the writer a unique vantage point to observe the candidate. Second, build the evidence: two or three specific examples, each following the situation-behavior-outcome structure. Not 'she showed leadership' but 'when our research team lost our lead RA three weeks before the submission deadline, she reorganized the division of work, held two additional team check-ins per week, and we submitted on time with our results intact.' Third, make a direct endorsement: name the specific program or role, state that you recommend without reservation, and connect the candidate's demonstrated qualities to what that program demands. Committees notice writers who make the connection explicit.

How AI helps busy professionals write specific letters

Most professionals asked to write a recommendation say yes intending to write something strong, then face a blank page and produce something generic because they do not remember the specific details they need. AI solves this by starting from structure and prompting you to fill in the evidence. Rather than writing a letter from scratch, you describe the relationship, paste the examples the candidate provided (always ask candidates to supply their own brag sheet), and ask AI to draft the letter. The output is typically 80% of the way there — the structure and language are polished, and you edit in the specific institutional knowledge and voice that only the writer can provide. This approach takes 15 minutes instead of 90 and produces a more compelling letter because the structure is sound from the start.

Step-by-step guide

Gather specific examples

Ask the candidate to provide 3 specific projects or moments where they excelled before you begin writing.

Establish your credibility

Open with your relationship to the candidate, how long you have known them, and in what capacity.

Build the evidence section

Ask AI to write one strong paragraph per example that follows: situation, behavior observed, outcome achieved.

Close with a direct endorsement

End with a specific statement about what role or program they would excel in and why you recommend them without reservation.

Ready-to-use prompts

Academic or professional recommendation letter

Act as a [WRITER'S TITLE, e.g. 'university professor' or 'senior manager']. Write a recommendation letter for [CANDIDATE NAME] applying to [PROGRAM/ROLE/COMPANY]. My relationship: [HOW LONG AND IN WHAT CAPACITY I KNOW THEM]. Their strongest qualities I observed directly: [2-3 SPECIFIC QUALITIES WITH CONTEXT]. Key examples to feature: (1) [SITUATION + WHAT THEY DID + RESULT]; (2) [SITUATION + WHAT THEY DID + RESULT]. What makes them exceptional vs their peers: [SPECIFIC COMPARISON]. Target length: [350-450 WORDS]. Format: formal letter with greeting, 3 substantive paragraphs, direct endorsement closing. Avoid generic phrases like 'hard worker', 'team player', or 'strong communication skills' unless supported by a specific example.

Why it works

The 'exceptional vs peers' input is the most important — it forces the writer to articulate comparative distinction, which is the single most persuasive element of any recommendation letter.

LinkedIn recommendation (short form)

Write a LinkedIn recommendation from [WRITER'S ROLE] to [CANDIDATE NAME], who worked as [CANDIDATE'S ROLE] for [DURATION]. The relationship: [HOW WRITER KNOWS THEM]. Two specific things to highlight: (1) [SPECIFIC ACHIEVEMENT OR BEHAVIOR WITH RESULT]; (2) [ANOTHER EXAMPLE]. The recommendation should read between 100 and 150 words, in first person, with a specific opening that does not start with 'I highly recommend', a middle section with one of the two examples in detail, and a closing endorsement for [TYPE OF ROLES OR COMPANIES CANDIDATE IS TARGETING]. No vague praise without backing.

Why it works

Short LinkedIn recommendations are read more than long ones, so the constraints on length and specificity push the output toward the most impactful evidence rather than comprehensive coverage.

Practical tips

✦Before writing a word, ask the candidate to send you their brag sheet: the specific projects you worked on together, measurable outcomes, and what they want the letter to emphasize. You cannot write specifically without these inputs.
✦Each example paragraph should follow situation-behavior-outcome: what was happening, what specifically the candidate did (not the team — the candidate), and what the measurable or observable result was.
✦The closing paragraph should name the specific program or role type and explain why the candidate's demonstrated qualities map directly to what that context demands. Generic closings waste the letter's strongest position.
✦Ask AI to flag any phrase in your draft that could describe any other strong candidate — those are the generic phrases that weaken the letter and should be replaced with something specific to this person.
✦If you genuinely cannot think of two strong, specific examples, decline to write the letter. A weak letter written out of obligation does the candidate more harm than no letter, especially in competitive programs.