# USAA — API Agent Score

> Score: 53/100 (Grade: D) | Domain: usaa.com | Rubric: 1.0.0 | Checked: November 3, 2026

USAA scored 53/100 (D), classified "not-ai-ready".

[View full report](https://www.postman.com/ai/ai-ready-apis/company/usaa)
[All organizations](https://www.postman.com/ai/ai-ready-apis/llms.txt)

## Summary

- Overall: 53/100 (D)
- Classification: not-ai-ready

## Category Scores

- API Design: 50/100 (D)
- Developer Experience: 100/100 (A+)
- Agent Discovery: 45/100 (F)
- Agent Understanding: 61/100 (C)
- Agent Usability: 40/100 (F)

## Check Results

### API Design

- [pass] Security & governance hygiene — Code is formatter-clean with zero linter errors and zero style warnings. Investigated: sdk 100%, wellknown 0%.
- [fail] Schema coverage & depth — no .d.ts found. Investigated: sdk 0%.
- [fail] Example coverage — No README found. Investigated: sdk 0%.
- [skip] Machine-readable, versioned contract — No surface produced evidence for this capability in this run.
- [skip] Auth declared & discoverable — No surface produced evidence for this capability in this run.

### Developer Experience

- [pass] Self-service developer portal — Signup page found at https://www.usaa.com/register?akredirect=true, but no free tier or sandbox is documented in docs/pricing, so an agent may hit a paywall. Investigated: docs 100%.
- [skip] Quickstart present — No surface produced evidence for this capability in this run.
- [skip] Code samples in docs — No surface produced evidence for this capability in this run.
- [skip] Description completeness — No surface produced evidence for this capability in this run.
- [skip] Changelog published — No surface produced evidence for this capability in this run.

### Agent Discovery

- [warn] Registry & SDK presence — No official SDKs were found on npm, PyPI, or GitHub, so agents must call the API by hand. Investigated: sdk 25%, docs 0%, cli 0%.
- [warn] Crawlable / AEO — Sitemap has 967 entries but no lastmod values, so crawlers can't tell what changed recently. Investigated: wellknown 50%.
- [fail] llms.txt present, valid & comprehensive — No llms-full.txt was found at any candidate origin, so agents can't grab your full docs in one fetch. Investigated: docs 0%.
- [skip] Docs reachable, not hard auth-gated — No surface produced evidence for this capability in this run.

### Agent Understanding

- [warn] Agent instructions file (AGENTS.md) — No agents.md or skill.md context file found (nice-to-have for agent operation). Investigated: sdk 50%, wellknown 0%.
- [skip] Machine-readable errors (RFC 9457) — No surface produced evidence for this capability in this run.
- [skip] Operation purpose clarity — No surface produced evidence for this capability in this run.
- [skip] Agent-navigable, token-efficient docs — No surface produced evidence for this capability in this run.
- [skip] Docs structured data — No surface produced evidence for this capability in this run.
- [skip] Description consistency across surfaces — Only 1 surface description(s) with ≥6 tokens available; need at least 2 to compare.

### Agent Usability

- [fail] Runnable collection with test scripts — No README found. Investigated: sdk 0%, platform 0%.
- [skip] Idempotency documented — No surface produced evidence for this capability in this run.
- [skip] Rate-limit signaling — No surface produced evidence for this capability in this run.
- [skip] Pagination documented & consistent — No surface produced evidence for this capability in this run.
- [skip] Sandbox separation — No surface produced evidence for this capability in this run.

## Executive Summary

USAA's API program has meaningful room to close before partners and AI agents can self-serve reliably. The strongest signal is in agent context, where baseline understanding is in place, but the program's most critical gaps are in discoverability and design completeness — meaning partners and their agents currently have no structured way to find, explore, or build against the API surface without direct engagement. Prioritize publishing a machine-readable discovery layer and runnable examples first, then lock in a typed, example-rich contract so partners can move from evaluation to integration without friction.