Skip to main content

Testing Guides by Industry

Short answer

These guides show Arrange / Act / Assert patterns for domain risk—fixture-backed setup, focused UI actions, and probe assertions—using TestChimp SmartTests, markdown plans, and per-PR /testchimp test.

Why domain guides matter

Industry risk is not generic “click through the app.” Checkout has coupons and payment state. Fintech has ledger invariants. SaaS has activation funnels. AI products have non-deterministic UI that still needs probe Assert.

Each guide below walks a concrete failure mode and shows how TestChimp applies:

  1. Arrange — seed routes / fixtures (parallel-safe)
  2. Act — focused Playwright UI steps
  3. Assert — probe endpoints for authoritative state

Canonical reference: expired-coupon pattern.

Ecommerce

GuideFailure modes covered
Checkout flowsExpired coupons, payment capture, parallel CI
Cart and promo codesStacking rules, run-scoped coupons, revenue leaks

Built with Lovable or Cursor? Combine with built-with guides.

Fintech

GuideFailure modes covered
Fintech web appsTransfers, limits, idempotency, ledger probes

SaaS & platform

GuideFailure modes covered
SaaS onboardingFunnels, screen states, activation
Flaky E2E in startupsShared data, world-state, fixture discipline

AI products

GuideFailure modes covered
AI-powered web appsHybrid ai.act for volatile UI + probes for truth

Regulated & vertical SaaS

GuideFailure modes covered
HR & applicant trackingPipeline stages, PII hygiene, referral paths
Healthcare patient portalsChart ACL, PHI in URLs, caregiver access
EdTech enrollmentPrerequisites, waitlists, progress gating
Insurance quote & bindRating knock-outs, multi-line wizards
Real estate listingsMap/list desync, geo filters, saved searches
Legal e-signaturesSigner roles, envelope state, parallel signing
CMS publishingScheduled publish, rollback, editorial states
Audit & compliance logsImmutable audit rows, actor attribution

Also see cross-cutting guides: Stripe, Firebase auth, GDPR.

How TestChimp applies to domain guides

LayerWhat it does
Test planning as codeDomain rules—coupon stacking, KYC tiers, enrollment DAGs—live in markdown plans agents read on every PR
Requirement traceability// @Scenario: links roll up which failure modes each SmartTest covers
ExploreChimpNon-functional bugs on high-traffic domain paths—slow checkout, confusing knock-out copy, misleading balances
Post-deploy evolve/testchimp evolve closes plan gaps after deploy; production signals (including TrueCoverage) inform priority—not replace scenarios

Record-replay and no-code tools rarely model domain matrices or probe Assert. TestChimp runs in Claude or Cursor with the TestChimp skill—see record-replay vs TestChimp.

How to use these guides

  1. Read the guide closest to your domain
  2. Copy Arrange/Act/Assert patterns into markdown scenarios
  3. Run /testchimp init if harness is not scaffolded yet
  4. Gate the next domain PR with /testchimp test
  5. After deploy, run /testchimp evolve when plan gaps or explore findings surface new risk

Note: Builder tools (Lovable, Cursor, etc.) live under Built-with guides—not industry verticals.

Frequently asked questions

Which industry guide should we read first?

Start with the expired-coupon Arrange/Act/Assert pattern, then open the guide closest to your domain—checkout, fintech ledger probes, SaaS onboarding funnels, or flaky E2E fixes for fast CI.

Can a small ecommerce or fintech team run this without QA hires?

Yes. Domain guides show seed routes and probe assertions so developers run `/testchimp test` on PRs—coupon collisions, transfer limits, and activation funnels stay covered via SmartTests linked to markdown scenarios.

Do vertical guides cover native mobile apps?

Web patterns dominate checkout and onboarding guides, but the same markdown plans and `// @Scenario:` model applies to Mobilewright on iOS and Android.

How do agents use these guides in practice?

Agents read scenario files and domain guides when `/testchimp test` runs—authoring SmartTests with fixture-backed Arrange and probe Assert instead of UI-only recordings that flake in parallel CI.