AgentSkillsCN

fintech-patterns

常见金融科技客户模式、异议与成功案例。合规处理、长对话管理与案例研究。

SKILL.md
--- frontmatter
name: fintech-patterns
description: Common fintech customer patterns, objections, and success stories. Compliance handling, long conversation management, and case studies.

Fintech Patterns Skill

When to Use

  • •Customer is in fintech/financial services
  • •Compliance concerns come up
  • •Questions about regulatory requirements
  • •Long conversation challenges (common in financial chat)
  • •Need case studies from similar customers

Common Fintech Patterns

Pattern 1: Compliance Concern + Pause

Signal: Customer pauses, then mentions compliance Example: "Um, one thing though — we're in fintech, so there's compliance stuff..." Response: Lead with SOC2/audit success stories, show how other fintechs solved it

Pattern 2: Token Cost + Scale Fear

Signal: Growing user base, worried about costs Example: "Our costs are exploding as we scale..." Response: Show ROI of Context Editing, give specific numbers

Pattern 3: "Claude Forgets"

Signal: Users complaining about lost context Example: "By message 15, Claude forgets what we discussed in message 3" Response: Context Editing with persistent facts pattern

Typical Fintech Requirements

  1. •Compliance: SOC2, GDPR, financial regulations
  2. •Audit trails: Logging all AI decisions
  3. •Data residency: Where data is processed
  4. •Long conversations: Users ask many follow-ups
  5. •Accuracy: Can't give wrong financial advice

Success Stories

Acme Wealth

  • •Problem: 40-50 turn conversations losing context
  • •Solution: Context Editing with persistent facts
  • •Result: Passed SOC2 audit, 65% token reduction

FinBot (reference customer)

  • •Problem: Token costs scaling with user growth
  • •Solution: Rolling summarization strategy
  • •Result: 70% cost reduction, better UX

Response Guidelines

  1. •Acknowledge the industry: "Fintech has unique challenges..."
  2. •Lead with compliance: Always address regulatory concerns first
  3. •Use case studies: Reference similar customers
  4. •Be specific: Give numbers, not generalities