Recommendation letters for EB-1A, NIW, and O-1: who should write them, what goes inside, and why do letters from colleagues not work?

An applicant brought ten letters. Nine were from people he had worked with: supervisors, colleagues, project partners, all well known in the field, all with warm words about his talent and his sense of responsibility. The tenth was from a professor in another country who had never seen him, but had used his method in the lab for two years and described that in two pages of figures. In an RFE (Request for Evidence), the officer called the nine "letters from interested parties, confirming professional qualities" and quoted the tenth as evidence of an original contribution. One letter of the ten did the work, and it was the only one the applicant had not wanted to ask for - it felt awkward to bother a stranger. This article is about how to make all of them work.

In brief

• Letters are evidence on the criteria, not a character reference. They have to contain facts the officer can check.
• What decides it is the author's independence: half the letters from experts who know the work from its results, not from working together.
• Five to eight letters, each aimed at its own criterion, in different words, with dates, names, and figures.
• O-1 also needs an advisory opinion (a consultation from a peer group) from a union or an association. That is a different document.

Why does the officer want the letters, and how does he read them?

The EB-1A and O-1 criteria require recognition in the field. Publications, awards, and judging are recognition recorded in documents. Recommendation letters are recognition recorded by people, and the officer reads them with one question: who is this person, and why should I believe them. The applicant's supervisor has an interest in the outcome, a colleague even more so, and their letters of support confirm professional qualities, not standing in the field. An independent expert, who gains nothing from the petition, is a different matter: an expert opinion letter saying that the applicant's method is used, and the result is cited, becomes evidence of the contribution and of the reputation.

The first rule follows. Letters are not divided into "strong" and "weak." They are divided into dependent and independent. Dependent ones are needed: they confirm the role, the duties, and the result inside the company. Independent ones decide the case: they confirm recognition. A petition needs both, in a proportion of about one to one, and each one aimed at its own criterion. How the officer reads each criterion is in "What are the U.S. talent visa criteria".

Who should write them?

Independent experts. A professor who applies the applicant's method; the head of a client company that uses the product; an editor of a trade publication that has written about the work; an organizer of a competition the applicant won or judged; an author who cites the work. What they share is this: they know the work from its results, and they are not tied to the applicant by employment, kinship, or friendship. Finding them is easier than it seems - through citations, adoptions of the work, conferences - and asking is not something to be embarrassed about: experts write these letters regularly.

Supervisors and clients. They confirm the critical role, the duties, the result, the salary - with dates, titles, and figures. Their letters of support are read as dependent, but without them the critical-role criterion does not close.

Colleagues and partners. The weakest category. One or two letters are acceptable if they contain facts nobody else has. Ten are dead weight.

Well-known people who do not know the applicant. A letter that says "I have heard he is a good specialist," from a celebrity, weighs nothing. The officer checks the author's connection to the applicant's work.

What is in a letter the officer accepts?

Five parts, each one specific.

Who the author is: the position, the organization, the achievements in the field, why the opinion carries weight - two or three sentences of facts, not titles.

How the author knows the applicant: which work, since which year, in what role - joint or independent. Honesty matters more than the impression. "We have not worked together. I know the work from the publications and from using the method in our lab" is stronger than a vague "I have known him for a long time."

What the applicant actually did: a specific achievement for a specific criterion - a method, a product, a role - with the name, the date, the figures, and where it is used. One letter covers one or two criteria, not all of them at once.

Why this is recognition at the level of the field: what it is compared with, who else uses it, how it changed the practice. This is the paragraph that is most often missing, and it is what separates evidence from a character reference.

The conclusion: a direct statement that, in the author's opinion, the applicant is among the small percentage at the top of the field - and a signature with contact details USCIS can use to check the letter.

The length is one and a half to two pages. The language is English, or another language with a certified translation. The organization's letterhead is desirable. A notary is not needed.

Seven mistakes that turn letters into dead weight

Identical letters. Ten texts with the same phrases - "outstanding," "an invaluable contribution," "one of the best" - and the officer reads them as one text written by the applicant. Each letter in the author's own words, with its own structure.

A character reference instead of facts. "Talented, responsible, driven" - none of that can be checked. A date, a name, and a figure can.

A retelling of the resume. A letter that lists the applicant's jobs adds nothing to what the officer has already seen.

No dates and no titles. In the review "The officer accepted one criterion of seven", the critical-role criterion was not accepted for exactly this reason: the officer could not tell who was writing, or for which period.

Every letter from one organization. Five letters from one company are one point of recognition, not five.

A letter "about everything." If the author confirms the role, the contribution, the salary, and the judging, the officer doubts that the author knows all of it from personal experience.

An author who cannot be found. USCIS checks letters selectively. A signature with no contact details, or with details nobody answers, discounts that letter and casts a shadow on the rest.

How do you organize it: a brief, a draft, and a signature

The process we use on every petition takes four to eight weeks, and it starts not with the letters but with a map of the criteria: which three or four we are building, and which letter closes each one. For each letter, a brief: the criterion, the facts that need confirming, and what the author knows personally. The draft is written from the brief and sent to the author with a request to rewrite it in their own words, to add, or to cut. Recommenders almost always add something the applicant did not know. Signature, letterhead, contact details, translation - and the letter goes into the petition next to the documents it supports: the professor's expert opinion letter next to the publications on how the method was adopted, the client's letter of support next to the contract and the results.

There are usually five to eight recommenders, half of them independent experts. If there are no independent experts at all, that is a signal not about the letters but about the profile: the outside trace that can be confirmed is not there yet, and it has to be built - publications, judging, talks - before anyone is asked for a letter.

The O-1 letter without which the petition is incomplete

O-1 has a document that is often confused with recommendation letters: an advisory opinion (a consultation from a peer group), a written opinion from a union or a professional association in the applicant's field that the applicant meets the category. The petitioner requests it before filing. If there is no organization in the field, that is documented separately. Without the advisory opinion, an O-1 petition is incomplete. Recommendation letters are still needed, as they are for EB-1A - on the criteria, with the same rules of independence and of facts. The category itself is in "The O-1 visa".

After the request, the applicant from the opening paragraph replaced six of the nine letters: he found three independent experts through citations of his articles and rewrote the supervisors' letters with dates and figures. The second time, the officer quoted four letters out of seven. The question he asked us at the first meeting - "do I really have to bother strangers" - is the question we now ask on consultations ourselves: if nobody can recommend you except the people you worked with, the petition is not ready yet. Which criteria you already have, and who can confirm them, is what the quiz below shows.