{
  "schema": "https://ai-atoms.com/schemas/skill-v1.json",
  "type": "skill",
  "id": "skill/customer-research",
  "version": "1.0.0",
  "name": "customer-research",
  "description": "Multi-source research on a customer question or topic with source attribution. Use when a customer asks something you need to look up, investigating whether a bug has been reported before, checking what was previously told to a specific account, or gathering background before drafting a response.",
  "system_prompt_fragment": "# /customer-research\n\n> If you see unfamiliar placeholders or need to check which tools are connected, see [CONNECTORS.md](../../CONNECTORS.md).\n\nMulti-source research on a customer question, product topic, or account-related inquiry. Synthesizes findings from all available sources with clear attribution and confidence scoring.\n\n## Usage\n\n```\n/customer-research <question or topic>\n```\n\n## Workflow\n\n### 1. Parse the Research Request\n\nIdentify what type of research is needed:\n- **Customer question**: Something a customer has asked that needs an answer (e.g., \"Does our product support SSO with Okta?\")\n- **Issue investigation**: Background on a reported problem (e.g., \"Has this bug been reported before? What's the known workaround?\")\n- **Account context**: History with a specific customer (e.g., \"What did we tell Acme Corp last time they asked about this?\")\n- **Topic research**: General topic relevant to support work (e.g., \"Best practices for webhook retry logic\")\n\nBefore searching, clarify what you're actually trying to find:\n- Is this a factual question with a definitive answer?\n- Is this a contextual question requiring multiple perspectives?\n- Is this an exploratory question where the scope is still being defined?\n- Who is the audience for the answer (internal team, customer, leadership)?\n\n### 2. Search Available Sources\n\nSearch systematically through the source tiers below, adapting to what is connected. Don't stop at the first result — cross-reference across sources.\n\n**Tier 1 — Official Internal Sources (highest confidence):**\n- ~~knowledge base (if connected): product docs, runbooks, FAQs, policy documents\n- ~~cloud storage: internal documents, specs, guides, past research\n- Product roadmap (internal-facing): feature timelines, priorities\n\n**Tier 2 — Organizational Context:**\n- ~~CRM notes: account notes, activity history, previous answers, opportunity details\n- ~~support platform (if connected): previous resolutions, known issues, workarounds\n- Meeting notes: previous discussions, decisions, commitments\n\n**Tier 3 — Team Communications:**\n- ~~chat: search for the topic in relevant channels; check if teammates have discussed or answered this before\n- ~~email: search for previous correspondence on this topic\n- Calendar notes: meeting agendas and post-meeting notes\n\n**Tier 4 — External Sources:**\n- Web search: official documentation, blog posts, community forums\n- Public knowledge bases, help centers, release notes\n- Third-party documentation: integration partners, complementary tools\n\n**Tier 5 — Inferred or Analogical (use when direct sources don't yield answers):**\n- Similar situations: how similar questions were handled before\n- Analogous customers: what worked for comparable accounts\n- General best practices: industry standards and norms\n\n### 3. Synthesize Findings\n\nCompile results into a structured research brief:\n\n```\n## Research: [Question/Topic]\n\n### Answer\n[Clear, direct answer to the question — lead with the bottom line]\n\n**Confidence:** [High / Medium / Low]\n[Explain what drives the confidence level]\n\n### Key Findings\n\n**From [Source 1]:**\n- [Finding with specific detail]\n- [Finding with specific detail]\n\n**From [Source 2]:**\n- [Finding with specific detail]\n\n### Context & Nuance\n[Any caveats, edge cases, or additional context that matters]\n\n### Sources\n1. [Source name/link] — [what it contributed]\n2. [Source name/link] — [what it contributed]\n3. [Source name/link] — [what it contributed]\n\n### Gaps & Unknowns\n- [What couldn't be confirmed]\n- [What might need verification from a subject matter expert]\n\n### Recommended Next Steps\n- [Action if the answer needs to go to a customer]\n- [Action if further research is needed]\n- [Who to consult for verification if needed]\n```\n\n### 4. Handle Insufficient Sources\n\nIf no connected sources yield results:\n\n- Perform web research on the topic\n- Ask the user for internal context:\n  - \"I couldn't find this in connected sources. Do you have internal docs or knowledge base articles about this?\"\n  - \"Has your team discussed this topic before? Any ~~chat channels I should check?\"\n  - \"Is there a subject matter expert who would know the answer?\"\n- Be transparent about limitations:\n  - \"This answer is based on web research only — please verify against your internal documentation before sharing with the customer.\"\n  - \"I found a possible answer but couldn't confirm it from an authoritative internal source.\"\n\n### 5. Customer-Facing Considerations\n\nIf the research is to answer a customer question:\n\n- Flag if the answer involves product roadmap, pricing, legal, or security topics that may need review\n- Note if the answer differs from what may have been communicated previously\n- Suggest appropriate caveats for the customer-facing response\n- Offer to draft the customer response: \"Want me to draft a response to the customer based on these findings?\"\n\n### 6. Knowledge Capture\n\nAfter research is complete, suggest capturing the knowledge:\n\n- \"Should I save these findings to your knowledge base for future reference?\"\n- \"Want me to create a FAQ entry based on this research?\"\n- \"This might be worth documenting — should I draft a runbook entry?\"\n\nThis helps build institutional knowledge and reduces duplicate research effort across the team.\n\n---\n\n## Source Prioritization and Confidence\n\n### Confidence by Source Tier\n\n| Tier | Source Type | Confidence | Notes |\n|------|-------------|------------|-------|\n| 1 | Official internal docs, KB, policies | **High** | Trust unless clearly outdated — check dates |\n| 2 | CRM, support tickets, meeting notes | **Medium-High** | May be subjective or incomplete |\n| 3 | Chat, email, calendar notes | **Medium** | Informal, may be out of context or speculative |\n| 4 | Web, forums, third-party docs | **Low-Medium** | May not reflect your specific situation |\n| 5 | Inference, analogies, best practices | **Low** | Clearly flag as inference, not fact |\n\n### Confidence Levels\n\nAlways assign and communicate a confidence level:\n\n**High Confidence:**\n- Answer confirmed by official documentation or authoritative source\n- Multiple sources corroborate the same answer\n- Information is current (verified within a reasonable timeframe)\n- \"I'm confident this is accurate based on [source].\"\n\n**Medium Confidence:**\n- Answer found in informal sources (chat, email) but not official docs\n- Single source without corroboration\n- Information may be slightly outdated but likely still valid\n- \"Based on [source], this appears to be the case, but I'd recommend confirming with [team/person].\"\n\n**Low Confidence:**\n- Answer is inferred from related information\n- Sources are outdated or potentially unreliable\n- Contradictory information found across sources\n- \"I wasn't able to find a definitive answer. Based on [context], my best assessment is [answer], but this should be verified before sharing with the customer.\"\n\n**Unable to Determine:**\n- No relevant information found in any source\n- Question requires specialized knowledge not available in sources\n- \"I couldn't find information about this. I recommend reaching out to [suggested expert/team] for a definitive answer.\"\n\n### Handling Contradictions\n\nWhen sources disagree:\n1. Note the contradiction explicitly\n2. Identify which source is more authoritative or more recent\n3. Present both perspectives with context\n4. Recommend how to resolve the discrepancy\n5. If going to a customer: use the most conservative/cautious answer until resolved\n\n## When to Escalate vs. Answer Directly\n\n### Answer Directly When:\n- Official documentation clearly addresses the question\n- Multiple reliable sources corroborate the answer\n- The question is factual and non-sensitive\n- The answer doesn't involve commitments, timelines, or pricing\n- You've answered similar questions before with confirmed accuracy\n\n### Escalate or Verify When:\n- The answer involves product roadmap commitments or timelines\n- Pricing, legal terms, or contract-specific questions\n- Security, compliance, or data handling questions\n- The answer could set a precedent or create expectations\n- You found contradictory information in sources\n- The question involves a specific customer's custom configuration\n- The answer requires specialized expertise you don't have\n- The customer is at risk and the wrong answer could exacerbate the situation\n\n### Escalation Path:\n1. **Subject matter expert**: For technical or domain-specific questions\n2. **Product team**: For roadmap, feature, or capability questions\n3. **Legal/compliance**: For terms, privacy, security, or regulatory questions\n4. **Billing/finance**: For pricing, invoice, or payment-related questions\n5. **Engineering**: For custom configurations, bugs, or technical root causes\n6. **Leadership**: For strategic decisions, exceptions, or high-stakes situations\n\n## Research Documentation for Team Knowledge Base\n\nAfter completing research, capture the knowledge for future use.\n\n### When to Document:\n- Question has come up before or likely will again\n- Research took significant effort to compile\n- Answer required synthesizing multiple sources\n- Answer corrects a common misunderstanding\n- Answer involves nuance that's easy to get wrong\n\n### Documentation Format:\n```\n## [Question/Topic]\n\n**Last Verified:** [date]\n**Confidence:** [level]\n\n### Answer\n[Clear, direct answer]\n\n### Details\n[Supporting detail, context, and nuance]\n\n### Sources\n[Where this information came from]\n\n### Related Questions\n[Other questions this might help answer]\n\n### Review Notes\n[When to re-verify, what might change this answer]\n```\n\n### Knowledge Base Hygiene:\n- Date-stamp all entries\n- Flag entries that reference specific product versions or features\n- Review and update entries quarterly\n- Archive entries that are no longer relevant\n- Tag entries for searchability (by topic, product area, customer segment)",
  "applicable_domains": [
    "support",
    "customer"
  ],
  "invocation": [
    "/customer-research",
    "/customer-research <question or topic>"
  ],
  "tags": [
    "customer-support",
    "anthropics",
    "knowledge-work"
  ],
  "authored_by": "anthropics",
  "source_url": "https://github.com/anthropics/knowledge-work-plugins/blob/main/customer-support/skills/customer-research/SKILL.md",
  "lifecycle": "stable",
  "category": "support",
  "provenance": {
    "source": "anthropics/knowledge-work-plugins",
    "source_url": "https://github.com/anthropics/knowledge-work-plugins/blob/main/customer-support/skills/customer-research/SKILL.md",
    "author": "Anthropic",
    "license": "Apache-2.0",
    "notes": "Imported by scripts/import-anthropic-skills.py."
  }
}