Visalaw AI Team
How Immigration Teams Are Using Visalaw AI on O-1 Petitions
O-1 petitions live or die on one thing: matching evidence to criteria, and keeping that mapping consistent across the support letter, expert letters, exhibits, and brief. The client sends a folder of mixed documents. Someone has to read every page, decide which criterion each piece supports, spot what's missing, and make sure everything filed tells the same story.
Done by hand, that takes hours. And inconsistencies still slip through.
Here's how immigration teams are using Visalaw AI at each stage of an O-1A or O-1B matter, with the actual prompts they run. In every case, the attorney makes the judgment calls. The AI does the reading, sorting, and cross-checking that eats the day.
1. Get a case snapshot before you start
Upload the full client folder, then ask for the lay of the land:
"Summarize this O-1 matter: who the beneficiary is, the petitioning employer, the role and start date, and a categorized inventory of every document uploaded. Note any prior USCIS approvals."
In seconds you have a one-page summary of the beneficiary, the employer, the role, and an organized inventory of 20-plus mixed files. That replaces the manual folder review most teams do before any real analysis begins.
2. Map evidence to the regulatory criteria
This is the core of the case. O-1A has eight possible criteria under 8 CFR 214.2(o)(3)(iii); you need three. O-1B has its own set. The prompt:
"Map the uploaded evidence to the eight O-1A regulatory criteria. For each criterion, list which specific documents support it, rate the strength (strong / moderate / weak / not met), and explain the reasoning. Flag any criterion where the evidence looks weaker than it first appears."
The output is a criteria matrix: which documents support which criterion, a strength rating, and reasoning for each. In a recent test case, the platform identified four defensible criteria (original contributions, authorship, critical role, and high salary) and correctly rated awards and memberships as too weak to claim. That's the analysis an attorney would otherwise build over hours of page-turning.
3. Flag the gaps while there's still time to fix them
The most valuable moment in the workflow is before filing, when a missing document can still be requested with one client email. Ask:
"Based on the criteria mapping, list the gaps and weaknesses in this petition. For each, explain why it's a problem and what additional evidence or correction would strengthen it before filing. Separate must-fix from would-help."
The kinds of problems this surfaces are the ones that draw RFEs:
Each of these, filed as-is, is an RFE waiting to happen. Caught early, each is a fixable evidence request.
4. Run a consistency check across the whole record
Cross-document inconsistencies are the errors that usually get caught too late, if at all. The prompt:
"Check the entire document set for internal inconsistencies: job titles, dates, citation counts, salary figures, and any individual who appears in more than one role. List each inconsistency with the conflicting sources."
In the same test case, this pass caught things a tired human reviewer misses: the beneficiary listed as "Director of Machine Learning Systems" in the employment agreement but "Head of Machine Learning" in the support letter, duplicate publication entries inflating the CV, and one expert who signed the peer consultation letter and also appeared as an independent expert elsewhere in the record. That last one matters. Reusing the same person in two roles invites an independence question from the adjudicator.
The output also suggested resolutions: pick one primary title, use one citation snapshot, present dual roles explicitly where they're real.
5. Pressure-test the criteria like an adjudicator would
Before committing to a filing strategy, run the record through a skeptical read:
"Pressure-test each criterion we plan to claim as an adjudicator would. Where is the evidence thin, conclusory, or reliant on the employer alone? What would an RFE most likely ask for?"
This is useful precisely because it's adversarial. It tells you where the record leans too hard on the employer's own statements and predicts what USCIS is likely to question.
6. Build the exhibit list and the letter outline
Once the strategy is set, the assembly work goes fast:
"Create a filing-ready exhibit list. Number each exhibit, give it a short descriptive title, and note which O-1 criterion or criteria it supports. Order exhibits by criterion, strongest first."
"Draft an outline for the O-1 petition letter argument. For each criterion we're claiming, write the heading, the key facts, and the exhibits cited. Only include criteria supported by the evidence. Mark anything that still needs attorney review."
The exhibit list and the letter outline come from the same document set and the same criteria mapping, so they point at the same story by construction. No blank-page start on the brief, and no exhibit list that drifted out of sync with the argument.
A few more worth keeping handy
Where the attorney fits
None of this replaces attorney judgment, and it shouldn't. The attorney decides which criteria to claim, reviews every letter, and signs off before anything is filed. What changes is where the hours go. Instead of spending them reading, sorting, and cross-checking, the team spends them on strategy and the calls only a lawyer can make.
An O-1 petition is an argument built from evidence. Visalaw AI reads the evidence, maps it, flags the weak spots, and keeps the record consistent, so you get to the argument faster and file with fewer surprises.
View all blogs
Subscribe below to receive tips for using AI in immigration law, product updates, and relevant news in immigration.
Memphis, TN
