Claude to draft a disability benefits appeal
Nunc used Claude Code to prepare an appeal after their mother's disability benefits and care allowance applications were rejected. The available evidence consisted of more than fifty scanned documents spanning about twenty five years, making chronology and document references a substantial part of the work.
Original prompt
Shared by the source author. See source notes for context.
Read .txt versions for text content. Reference official disability regulations for legal basis.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- nunc
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author submitted a seven section appeal with more than thirty dated references; the second level decision was pending.
- Human involvement
- Nunc organized the evidence and checked every fact, citation and judgment before submitting the appeal.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Create text companions for the scanned medical records through an OCR pipeline.
- Give Claude project context and instructions to consult official regulations.
- Iterate the appeal, verify its references and submit reviewed Word or PDF documents.
Nunc used Claude Code to prepare an appeal after their mother's disability benefits and care allowance applications were rejected. The available evidence consisted of more than fifty scanned documents spanning about twenty five years, making chronology and document references a substantial part of the work.
A Python OCR pipeline using Gemini created text companions for the scanned images. A project instruction file told Claude to read those text versions and consult official disability regulations for the legal basis. That gave the drafting stage a searchable record collection rather than relying on a single summary of the family's situation.
The appeal went through eleven versions. Nunc reported a final document of roughly 230 lines, divided into seven sections with more than thirty dated references. Word and PDF outputs were prepared for submission. The author organized the scans, supplied context and checked every factual assertion, citation and judgment used in the argument.
The second level appeal had been submitted when the account was published, and its decision was still pending. No benefit award or recovered payment is reported. This case supports using an agent to organize evidence and draft a reviewed appeal; it does not establish eligibility, validate a legal argument independently or show that the authority accepted the requested outcome.
Prompt
Original prompt
Source: hellonunc.com · nunc
Frequently asked questions
Why create text companions for scans?
They made decades of scanned records readable and searchable during the drafting process.
What did the author verify?
Nunc checked factual assertions, citations and judgments in the appeal before submission.
Were benefits awarded?
No award is reported; the submitted second level appeal was awaiting a decision.
More details
About this account
Claude Code read OCR text, consulted regulations and iterated a structured appeal with dated document references.
Firsthand source opened and read. Prompt is a verbatim excerpt of the published original instruction.
Credit: nunc
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · Gemini · Python · Word · PDF
Benefits appeal · Medical scans · OCR · Evidence citations
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.