Help Scout Prompts

You are a support coach for [COMPANY]. Write an agent copilot playbook for the Help Scout shared inbox that shows when to use AI Drafts to start a reply from past conversations, when to use AI Assist to fix tone and translate, and when to rely on AI Summarize to catch up on long threads. Include a checklist for reviewing every draft before sending, and rules for handling collision detection alerts when another agent is already active on [INBOX].

Build a knowledge base tuning plan for [COMPANY] so Help Scout AI Answers resolves more questions from Docs and added sources without human handoffs. Include steps to find unanswered questions from Beacon chat logs, rewrite articles for direct answer extraction, add missing FAQs about [PRODUCT], set AI guardrails for topics that must go to humans, and test answers in the AI Agent Test tab before publishing changes.

Create a deflection content strategy for the Beacon help widget at [COMPANY] aimed at reducing inbound email volume for [TOP_QUESTION]. Cover which Docs articles to feature in Beacon search, how to write concise self-serve answers for AI Answers, where to place proactive in-app messages, and how to measure deflected conversations versus handoffs in Beacon engagement reports.

Design a workflow automation blueprint for the Help Scout shared inbox of [COMPANY] covering round robin assignment across [TEAM], tag rules that flag conversations from the [VIP_DOMAIN] domain, auto-routing of refund requests to the billing queue, and time-based escalations that respect SLA policies. Note where each workflow fires and how agents confirm no duplicate assignment through collision detection.

Create a QA scorecard template for [COMPANY] support conversations in Help Scout covering greeting quality, accuracy of the solution, tone before and after AI Assist edits, use of saved replies, and correct tagging. Define a 1 to 5 scale per criterion, a pass threshold for [TEAM], and a weekly review process that uses AI Summarize to speed up reading long threads during [MONTH] audits.

Write a CSAT recovery playbook for [COMPANY] after a customer leaves a rating of [LOW_RATING] or below in Help Scout. Include a saved reply template that apologizes and asks what went wrong, escalation steps for assigning the follow-up owner in the shared inbox, rules for when AI Drafts should and should not generate the apology, and a tracking plan to confirm the customer was re-contacted within one business day.

Generate a library of five saved reply templates for [COMPANY] support covering the most common [ISSUE_TYPE] requests. Each template needs a clear subject line, placeholder fields for the customer name and [TICKET_ID], links to the relevant Docs article, and a fallback sentence that routes complex cases to a specialist. Add notes on when agents should run AI Assist for tone before sending.

Outline a monthly analytics report for [COMPANY] using Help Scout reports for [MONTH], covering conversation volume per [INBOX], AI Answers resolution rate and handoff rate, average first reply time, CSAT trend, and tag distribution. Specify which advanced reports to pull, how to compare results against the previous month, and three concrete actions the team takes based on the numbers.

Design an AI Answers training and guardrail plan for [COMPANY] covering [PRODUCT]. List the intents to enable, the Docs articles and sources each intent draws from, guardrail topics that must escalate to a human, sample test questions for the AI Agent Test tab, and approval criteria for letting AI Answers go live in Beacon while keeping it within its read only permissions.

Write a two week onboarding playbook for new agent [NEW_AGENT] joining [COMPANY] support on Help Scout. Week one covers shared inbox basics, collision detection, saved replies, tags, and snooze and send later. Week two covers Docs and Beacon self service, AI Drafts and AI Assist usage rules, the workflows they will trigger, and a graded practice set of [PRACTICE_TICKET_COUNT] sample conversations with review checkpoints.