What a letter can and cannot do
USCIS policy states that letters of recommendation, on their own, are not persuasive evidence of extraordinary ability or of a nationally important endeavor; they are considered together with the documentary record. That is the right frame for writing them. A letter cannot create an achievement, but it can explain one: what a paper changed, why a method was adopted, how a role was critical, what a problem means to a field. Letters that do that work are read; letters that repeat the CV with adjectives are discounted, and a set of letters that repeat each other can actively hurt a petition by suggesting a template.
Choosing the writers
Four to six letters is typical. The set should include people with different relationships to the work: at least two or three independent experts who have never worked with the applicant but have used, cited or built on the work; a senior figure who can speak to the applicant's standing in the field; and, for NIW and for industry applicants, someone from industry, government or a user organization who can describe practical adoption. Supervisors and co-authors can write, but a set composed only of collaborators is weak, because every writer has a stake in the outcome. Prestige helps less than relevance: a well-known scientist in an adjacent field who knows the work only from the CV adds little.
What each letter should contain
Each letter should open with who the writer is and why their judgment is informed—position, field, and how they came to know the applicant's work, including that they have no personal relationship if that is the case. It should then describe one or two specific contributions in concrete terms: what the applicant did, what was new about it, and what happened as a result. The most useful sentences are the ones only this writer can supply: "My group adopted the applicant's calibration method in 2023 and it reduced our validation time by roughly half." For NIW letters, the writer should address why the proposed endeavor matters beyond the applicant's employer, in the writer's own domain. For EB-1A, the letter should map to the criteria the petition claims. Where the writer refers to a paper, a product or a dataset, the petition should include it as an exhibit.
Language to avoid
Hedged praise—"one of the more promising researchers I have encountered"—reads as faint praise to an officer. Superlatives without facts—"world-renowned", "a pioneer"—are ignored unless something in the record supports them. Paragraphs that could be about anyone in the field, legal conclusions ("clearly meets the criteria"), and discussions of immigration policy belong to the petition letter, not to a recommender. Letters should not describe the applicant's work in ways the CV and exhibits do not support, and should not disagree with each other on dates, titles or the description of the endeavor. Officers compare them.
The drafting process
In practice, the applicant or counsel prepares a draft for most writers, and the writer edits, adds and signs. That is accepted practice, but the letters must end up in different voices, with different emphases, reflecting what each writer actually knows. A writer should be sent the relevant exhibits—the papers, the citation data, the product documentation—so the letter is accurate, and should be asked to add the specifics only they have. The final letter goes on institutional letterhead where possible, with the writer's title, contact details and signature, and the writer should be willing to confirm it if asked. Keep the signed originals; USCIS occasionally verifies.
Common mistakes
Every letter from a co-author or former supervisor. Six letters with the same structure and the same three examples. A letter praising work the petition does not mention, or describing the endeavor differently from the petition. A letter from a writer whose own credentials are not documented. A letter that contradicts the CV on a date or a title. A letter signed by a writer who, when contacted, does not remember it. Each of these is avoidable, and each appears regularly in requests for evidence.
References
How Lumen Law & Co. can help
Recommendation letters should explain specific contributions from credible, genuinely informed perspectives. We plan the letters around the case record and applicable adjudication guidance.
- Distinguish firsthand factual witnesses, independent recommenders, and non-independent recommenders and assess the role each can play.
- Develop a balanced recommender mix based on the achievements and evidentiary gaps in the case.
- Identify verifiable facts and draft tailored letters for each recommender’s independent review and approval.
- Check consistency with research, projects, citations, positions, and the rest of the filing.