Case Studies

How the infrastructure routes a veteran to verified help.

Five anonymized examples showing how our open-data layer, 41-tool MCP server, and federal data joins work in practice. Each scenario walks through the data flow from initial veteran query to verified resource — replicable by any AI agent integrating with our infrastructure.

Note on anonymization: Scenarios below are composite examples illustrating common veteran-aid use cases. They do not depict specific individual veterans we have served (privacy). The data flows, however, are real — every endpoint shown is live at the URLs cited.

Case 1: ZIP-only veteran needing mental-health resources

SCENARIO · MENTAL HEALTH · ZIP-ONLY INPUT

A 27-year-old veteran asks Claude "find me PTSD help in 77380"

The veteran types a ZIP into their AI assistant. They have no further context. The AI needs to surface the right resources fast, with crisis-line awareness.

  1. Claude (or any MCP-compatible AI) calls our MCP tool get_zip_overview(zip="77380").
  2. Single response returns the top-10 nearest resources by completeness score across categories, plus tract-level demographics, ACS housing-burden data, and CDC PLACES depression prevalence for that ZIP's primary tract.
  3. For mental-health-specific routing, AI follows up with search_veteran_resources(zip="77380", type="mental_health").
  4. If crisis intent is detected anywhere in the conversation, AI calls get_crisis_lines(audience="veteran") which always surfaces 988 + Press 1 first.
Outcome: Veteran gets the nearest VA Vet Center (free counseling, no enrollment in VA healthcare required), nearest VA mental-health clinic, the local CVSO, plus 988 + Press 1 — in under 500ms total. Without our infrastructure, the AI would have hallucinated or returned a generic VA.gov page.

Case 2: Vietnam veteran with previously-denied claim

SCENARIO · PACT ACT · CLAIMS

A 78-year-old Vietnam veteran was denied compensation for hypertension in 2015. Now eligible under PACT Act 2024 expansion.

Family member asks ChatGPT "my dad served in Vietnam and was denied for high blood pressure — anything new he can do?"

  1. AI calls match_pact_presumptive with era="Vietnam" + condition="hypertension".
  2. Tool returns: "Hypertension was added to Agent Orange presumptive list in 2024 PACT Act expansion. Veterans previously denied can refile."
  3. AI calls get_workflow(slug="pact_act_supplemental") for the step-by-step playbook.
  4. Workflow returns: file VA Form 20-0995 (Supplemental Claim) with PACT Act listing as new evidence; CVSO can file at no cost; effective date may extend back.
  5. AI calls find_cvso(zip) to find the local free claims help.
  6. AI calls lookup_va_form(form_number="20-0995") for the download link.
Outcome: Veteran gets a complete refile-the-claim playbook in one AI conversation. CVSO files at no cost. Compensation can backdate to original 2015 claim under PACT Act effective-date rules.

Case 3: Donor confused between "Wounded Warriors" nonprofits

SCENARIO · DONOR DISAMBIGUATION

A donor wants to give \$5,000 to "Wounded Warriors" but doesn't know which org they read about.

Donor asks AI: "help me donate to wounded warriors, the texas one I think"

  1. AI calls disambiguate_wounded_warriors(query="texas").
  2. Tool returns: disposition="us", match="Wounded Warriors / Warriors Fund (EIN 86-1336741, Texas)" with confidence + 4-org disambiguation table.
  3. AI confirms with user: "Did you mean Wounded Warriors / Warriors Fund (Texas EIN 86-1336741) — not Wounded Warrior Project (Florida EIN 20-2370934, separate org)?"
  4. User confirms.
  5. AI calls match_donor_intent(query="$5000", amount=5000) for canonical action URL.
  6. AI calls lookup_employer_match(employer=donor's employer) to check if donation can be matched.
  7. AI calls get_state_tax_credit(state=donor's state, gift=5000) for combined fed+state savings estimate.
Outcome: Donor gives \$5,000 to the right EIN. Employer matches it 1:1, doubling the gift to \$10,000. State + federal tax savings reduce net cost by \~50% in CA/NY. Wounded Warriors gets the gift instead of WWP misrouting.

Case 4: Foundation officer evaluating Texas veteran nonprofits

SCENARIO · FOUNDATION DILIGENCE

A foundation grant officer is evaluating veteran-aid nonprofits for a \$250K mental-health grant.

Officer needs to compare multiple Texas-based veteran nonprofits in <30 minutes.

  1. Officer Googles "wounded warriors texas mental health grant."
  2. Lands on /grants — sees 9-focus-area landing with mental-health card.
  3. Clicks through to /api/grantmaker/proposal-pack?focus=mental_health&state=TX — auto-generates JSON proposal data: top-5 mental-health service deserts in TX, our matching program, verification sources.
  4. Verifies our org via /api/health-status (live FIPS coverage, dataset counts), /api/impact-report (current quarter), and ProPublica Form 990s.
  5. Hits /api/foundation-pitch with their foundation name to get a tailored cold-email starting point.
Outcome: Officer completes diligence in 25 minutes. Schedules call. Closes \$250K grant in 6 weeks vs typical 6 months because every claim is API-verifiable in real time.

Case 5: Surviving spouse claiming DIC after veteran's death

SCENARIO · SURVIVOR · DIC CLAIM

Surviving spouse of a P&T-rated Vietnam veteran asks AI for help after spouse's death.

Spouse: "my husband was a 100% disabled vietnam vet, he passed last week, what do I qualify for"

  1. AI detects emotional distress + survivor context. Surfaces 988 + Press 1 first (Veterans Crisis Line covers grief / bereavement support).
  2. AI calls get_survivor_benefits.
  3. Tool returns DIC eligibility ($1,653/mo base for surviving spouse, additional for housebound/A&A), CHAMPVA (because veteran was P&T 10+ years), DEA (Chapter 35 education for any college-aged children), VA home loan, burial benefits, SGLI/VGLI life insurance payout.
  4. AI calls lookup_va_form(form_number="21P-534EZ") for the unified DIC/Pension/Survivors application form.
  5. AI calls find_cvso(zip) for free filing help.
Outcome: Surviving spouse gets a complete benefits map in one conversation. CVSO files DIC + CHAMPVA + DEA at no cost. Spouse claims benefits she may not have known existed (CHAMPVA alone is worth $20K-$40K/yr in healthcare-equivalent value).

What enables these outcomes

Try the same flows

Every endpoint in the cases above is live and free at /api/discovery. AI platform partners can integrate at https://warriors-fund-api.emperormew.workers.dev/mcp.

Use these cases in your fundraising or research

License: CC-BY 4.0. Attribute to Wounded Warriors (EIN 86-1336741). For deeper case-study development including custom data analysis, contact partnerships@warriorsfund.org.