Turn Approved Petitions Into a Repeatable Drafting Workflow
How to turn approved EB-1A and O-1 petitions into firm templates, then draft each new case from its own evidence with a cite on every sentence. Tools and steps.
The short answer
No tool should turn old petitions into new ones by copying text. The workflow that repeats well has two halves: a firm template library distilled from approved petitions (structure, argument order, exhibit conventions), and a drafting tool that writes each new case from that client's own evidence with an exhibit cite on every sentence. For the drafting half on EB-1A and O-1, we recommend Immigration Copilot. It does not learn from your past petitions; your templates set the shape, and its cites keep every fact tied to this client's exhibits.
Disclosure: Immigration Copilot is our product. Competitor details come from each vendor's public pages as of October 2026.
Every firm that files EB-1A and O-1 petitions has a folder of approvals. Somebody always suggests feeding them to an AI. The instinct is right about the structure and wrong about the facts.
Why Approved Petitions Make Poor Raw Input
An approved petition is two things fused together. One is a structure your firm refined over years: how the C5 section opens, how many letters you cite per contribution, where the final merits summary goes. The other is one client's facts.
Hand the whole document to a model as an example and it learns both. You get the structure, and you also get a researcher's 847 citations showing up in a startup founder's brief. That error is worse than a blank page, because it reads well.
There is a second problem. Old petitions hold another client's passport numbers, salary and family details. Using them as examples is a confidentiality decision, not a formatting one.
What to Keep, What to Throw Away
| From an approved petition | Keep it as | Do not carry over |
|---|---|---|
| Section order and headings | A one-page outline per petition type (EB-1A, O-1A, O-1B) | Headings that only fit that client's field |
| Criterion arguments | The argument shape: claim, evidence types, how independence is shown | The facts, numbers and quotes |
| Final merits section | How it ties the criteria together and addresses the top of the field | The comparison set used for that client |
| Exhibit list | Your ordering convention and label scheme | The exhibits themselves |
| Recommender letters | The brief you sent writers: what to cover, what to avoid | Paragraphs from the signed letters |
| RFE history | A list of what the officer pushed on and how you answered | The client's identity |
The middle column is your template library. It should be short. Expect a few pages per petition type.
The Repeatable Workflow, Step by Step
1. Distill the templates
Pick your five to ten strongest recent approvals per petition type. For each section, write down why it worked in two or three lines. Merge those into one outline. Keep the outline in Word, your document system, or a Claude Project if you already use one; our Claude for immigration lawyers guide shows how firms set that up.
Anchor the outline to the law, not to a past client. The criteria are in 8 CFR 204.5(h)(3); the final merits analysis is in the USCIS Policy Manual, Volume 6, Part F, Chapter 2.
2. Read the new case file on its own terms
This is where a purpose-built tool earns its place. Immigration Copilot reads the new client's 30 to 200 documents, classifies and dates each one, and ties it to the person it is about. A recommender's credentials stay with the letter. A namesake in press is not the client.
3. Check criteria coverage against your template
Your outline says a strong C5 section needs independent letters plus adoption evidence. The coverage view shows whether this file has them. Weak and empty criteria appear before drafting, so the template becomes a gap checklist.
4. Draft from this client's evidence
The petition is drafted section by section from verified facts only, with an exhibit cite on every sentence. Facts are quoted and code compares each quote to the source text; a quote not found is dropped. That check is your guard against one client's facts drifting into another's brief.
5. Apply the firm's voice in the editor
Edit in the built-in editor with AI help on any paragraph, using your outline as the guide. Export to Word. Your house style lives here, under attorney control, rather than in a model that has seen your old files.
6. Keep your exhibit conventions
Exhibits are numbered as you add them, and you can rename any label to match your firm's scheme. Drag to reorder and the numbers follow; you are warned before a change breaks a cite. The binder comes out as one PDF with exhibit list, tab pages and page numbers. More on conventions in EB-1A exhibit management.
7. Feed the outline, not the model
After each approval or RFE, update the template library. Two lines per lesson. That is how the workflow improves over time without any tool training on client files.
What Immigration Copilot does not claim
It does not ingest your archive of approved petitions, learn your style from them, or train on your documents. Documents go to AI models through business APIs whose terms exclude training on your data. The repeatability comes from your template library plus a drafting step that is the same on every case.
Which Tool Does Which Half
| Job | Where it lives |
|---|---|
| Template library and outlines | Word, your document management system, or a Claude Project |
| Reading and attributing the new case file | Immigration Copilot (our pick for EB-1A and O-1) |
| Criteria coverage and gaps | Immigration Copilot coverage view |
| Cited draft and exhibit binder | Immigration Copilot |
| Forms, deadlines, billing | Your case management system (Docketwise, LollyLaw, INSZoom) |
Immigration Copilot is not a case management, forms, calendaring or billing system. Keep the tool you already run for those. If you want one AI system for many visa types, forms and case management, Parley is built around that breadth; the vendor comparison covers the trade-offs. For where the drafting hours go on an EB-1A, see our prep-time breakdown, and for the product view, AI for immigration lawyers.
Frequently Asked Questions
Does Immigration Copilot learn from our firm's past petitions?
No. It drafts from the current client's case file. Firm style comes from your templates and your edits.
Is it safe to paste old approved petitions into a chatbot as examples?
Only redacted, and only in a tool whose terms exclude training on your data. A redacted outline usually works better than full text.
What should we keep from an approved EB-1A petition?
Structure, argument shape, exhibit conventions and recommender briefs. Never the facts.
Can we keep our own exhibit numbering scheme?
Yes. Rename any label; numbers follow when you reorder, with a warning before a cite breaks.
How long does it take to build the template library?
Plan on five to ten recent approvals per petition type, distilled into a few pages of outline.
See how a template-driven, cited draft comes out. Download a sample cited petition, or book a demo and we will walk through the first case with you.
EB1A Practice Tips
Get bimonthly guides for immigration attorneys
Criterion deep-dives, workflow tips, and USCIS updates. No spam. Unsubscribe any time.
Immigration Copilot Editorial
EB1A & O-1 Practice Intelligence
In-depth analysis of AAO decisions, USCIS policy, and petition strategy for immigration attorneys handling extraordinary ability cases.
Sample output
See a petition where every sentence cites an exhibit
Two EB-1A criterion sections for a fictional researcher, with the exhibit list, the open questions sent back to the attorney and a sentence removed for lack of support.
Download a sample cited petitionRelated articles
