菜单
Find the high-volume customer intents your AI Agent has no Guidance for, then draft Guidance to close each gap. Use to systematically raise AI Agent coverage by aiming new knowledge at what customers actually ask most.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Build a targeted win-back segment from support signal — customers who hit a problem, left low CSAT, or went quiet after a bad experience — as a list you can hand to your email/SMS tool. The most qualified re-engagement audience you have is in your tickets. Read-only.
Turn the last 7 days of Gorgias tickets into an executive-ready weekly CX digest — analyzed across five business functions (Product, Marketing, Ecommerce/Revenue, Product Marketing, Leadership) with a TLDR, signal-status tagging, a PDP-gap table, and owner-tagged actions. Read-only; optionally publishes to Notion.
Build a back-in-stock demand list from customers who asked about out-of-stock or discontinued products — ranked by demand, with a per-product notify list — to drive back-in-stock campaigns and inform the next buy. Read-only.
Turn recurring customer language, objections, and praise from tickets into marketing angles, FAQ content, and on-site or ad copy — written in your customers' own words. The best copy you'll write is already in your inbox. Read-only.
Find the products generating the most pre-sales questions and post-purchase confusion, and get a concrete list of what to add to each product page. Turns support tickets into PDP improvements that lift conversion and cut returns. Read-only.
Cluster why customers actually return products — sizing, quality, expectation gaps — by product, from ticket and return language, so you can fix the root cause instead of just processing the refund. Returns are a margin killer; this finds where they start. Read-only.
| name | intent-gap-guidance |
| description | Find the high-volume customer intents your AI Agent has no Guidance for, then draft Guidance to close each gap. Use to systematically raise AI Agent coverage by aiming new knowledge at what customers actually ask most. |
The fastest way to raise AI Agent coverage isn't writing more Guidance — it's writing the right Guidance. Every gap you close is repetitive volume your team never has to touch again. This recipe finds the topics customers ask about most that your AI Agent currently has no answer for, and drafts Guidance for each, ranked by impact. Drafts only; you review and publish.
weekly-voc-digest surfaces a theme with no AI answer| Variable | Example |
|---|---|
{{MIN_VOLUME}} | "at least 20 tickets in the last 90 days" |
{{TOP_N}} | "the top 5 gaps" — how many to draft this round |
{{WINDOW}} | "the last 90 days" |
Using the Gorgias MCP, help me close my Gorgias AI Agent's biggest knowledge gaps.
- Look at my ticket intents over {{WINDOW}}. Identify high-volume topics ({{MIN_VOLUME}}) where my AI Agent has no matching Guidance, or where it's handing these over instead of resolving them.
- Rank the gaps by volume — biggest impact first.
- For {{TOP_N}} gaps, draft a Guidance for each. Ground each draft in how my team actually answers these tickets today (use real resolved tickets as the source of truth, not generic advice).
- Create them as drafts. List what you created so I can review and publish.
{{TOP_N}} — steady gains beat one giant batch nobody reviews.