Find supporting pages in VA records for a disability claim
A veteran used Gemini to scan about 9,000 pages of VA records and pinpoint the pages and dates supporting their claim, which was decided in about a month. Task: Suggested implementation: A veteran downloads Blue Button records from va.gov and uploads approximately 9,000 pages to Google Gemini. The AI scans the documents to identify claimable conditions, surfacing specific pages, treatment dates, and examples like chiropractor visits. The veteran reviews these AI-identified details to verify accuracy before filing the claim, replacing manual page-by-page searching.; tools: Gemini. Human role: The veteran manually downloaded Blue Button records from va.gov, uploaded them to Gemini, and reviewed the AI-identified pages and dates to verify accuracy before filing the claim.; limitations: Handling very large records (e.g., 9,000 pages) requires tools with large context windows, such as Gemini, though specific costs are not disclosed. The source is a Reddit post with secondary evidence reliability, and the original link returned a 403 error, limiting direct verification. Human approval is critical to ensure AI-identified dates and treatments match original documents before submission..
Tools
Published on Agent Work Atlas:
Overview
A veteran used Gemini to scan about 9,000 pages of VA records and pinpoint the pages and dates supporting their claim, which was decided in about a month.
Workflow
Suggested implementation: A veteran downloads Blue Button records from va.gov and uploads approximately 9,000 pages to Google Gemini. The AI scans the documents to identify claimable conditions, surfacing specific pages, treatment dates, and examples like chiropractor visits. The veteran reviews these AI-identified details to verify accuracy before filing the claim, replacing manual page-by-page searching.
Reported result
The author reports their own claim was decided in about one month from filing to decision letter. They also state they helped another veteran reach a 70% rating in two months using this method. These outcomes are self-reported and lack independent verification or detailed metrics on time saved versus manual review.
Human role
The veteran manually downloaded Blue Button records from va.gov, uploaded them to Gemini, and reviewed the AI-identified pages and dates to verify accuracy before filing the claim.
Limitations
Handling very large records (e.g., 9,000 pages) requires tools with large context windows, such as Gemini, though specific costs are not disclosed. The source is a Reddit post with secondary evidence reliability, and the original link returned a 403 error, limiting direct verification. Human approval is critical to ensure AI-identified dates and treatments match original documents before submission.
Try it
Act as a VA disability claim assistant. I will provide a PDF of my VA Blue Button records. First, perform a dry run on the first 50 pages to identify potential claimable conditions, listing the page number, treatment date, and provider for each finding. Wait for my confirmation that the extraction logic is accurate. Once confirmed, process the remaining pages. Output a structured table with columns: 'Condition', 'Supporting Page', 'Date', 'Treatment Type', and 'Quote'. Do not file any claims or spend money. If the file is too large to process at once, ask me to split it. Verify all dates against the original text before finalizing the list.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.