Unseen Lifeline
An accessibility-centered human-AI workflow for turning an urgent institutional notice into verified requirements, evidence needs, deadlines, communications, and a sequenced response plan.
The problem
People can lose rights, services, or opportunities because a process is difficult to navigate.
High-stakes notices often combine dense language, strict deadlines, fragmented documentation, multiple agencies, and severe cognitive load. The failure may be procedural rather than a lack of underlying eligibility or merit.
Unseen Lifeline is designed to reduce that burden by separating source evidence from interpretation, making deadlines and missing information visible, preparing draft communications, and keeping final review and submission with the person or their authorized support.
Representative transformation
From an overwhelming notice to a reviewable response plan.
The fictional example below demonstrates the structure without exposing personal records or representing legal advice.
“Provide the requested work-activity forms and supporting records within 10 days.”
The notice includes multiple forms, a response deadline, unfamiliar terminology, possible benefit consequences, and no organized list of what evidence already exists.
One deadline, four evidence needs, three open questions, and a sequenced action plan.
The workflow extracts the stated requirement, records the source language, identifies missing records, prepares questions for the agency, drafts a good-cause explanation for review, and creates a submission checklist.
| Prepared item | Example output | Human review required |
|---|---|---|
| Deadline | Response due: August 16, 2026, based on the stated 10-day period | Confirm receipt date and governing calculation |
| Evidence checklist | Forms, wage records, employer dates, hospitalization documentation | Confirm relevance, completeness, and privacy |
| Open questions | Which months are covered? Is an extension available? Where should records be submitted? | Contact the agency or authorized advisor |
| Draft communication | Request clarification and explain documented barriers affecting timely response | Edit, approve, and send personally or through authorized support |
| Action sequence | Verify deadline → gather records → complete drafts → review → submit → preserve confirmation | Choose actions and confirm completion |
Workflow architecture
Source first, then requirements, evidence, review, and action.
The process is designed to preserve uncertainty and prevent a draft from being mistaken for an approved submission.
Preserve source
Retain the original notice and identify the issuing organization, date, and stated consequence.
Decompose
Separate requirements, deadlines, forms, evidence requests, contacts, and unclear terms.
Compare evidence
Mark what exists, what is missing, what conflicts, and what requires verification.
Prepare support
Draft questions, communication, accommodation requests, checklists, and an action sequence.
Human decision
Review accuracy, obtain advice where needed, approve the final response, submit, and retain proof.
Evidence artifacts
What the capstone and case study demonstrate.
The project is presented as an evaluated concept and structured workflow, not as a deployed autonomous benefits, legal, medical, or eligibility system.
Requirement decomposition
Source-linked table of requirements, dates, forms, contacts, consequences, and ambiguous language.
Missing-information checklist
A review queue that distinguishes available evidence, missing records, conflicts, and verification needs.
Deadline and consequence map
Visible dates, dependencies, severity, reversibility, and escalation points.
Draft support
Questions, clarification requests, accommodation language, and response drafts tied to source facts.
Sequenced action plan
Tasks ordered by deadline, dependency, effort, and consequence, with completion evidence.
Human review gates
Explicit checkpoints before advice, submission, disclosure, or consequential action.
Selected design decisions
Accessibility is part of the operating architecture.
- Preserve the exact source language before simplifying it.
- Separate confirmed facts, inferred meaning, open questions, and draft language.
- Make the smallest next action visible instead of presenting one large response task.
- Track deadlines with their source and calculation assumptions.
- Return uncertainty to the user rather than silently resolving it.
- Keep legal, medical, financial, benefits, and eligibility judgment with qualified people.
Current maturity and boundaries
Unseen Lifeline is an evaluated capstone, product concept, and structured case study. It is not a deployed autonomous platform and does not determine eligibility, provide professional advice, or submit information on a person's behalf. Any real use would require source verification, privacy controls, qualified review where appropriate, and explicit user approval.