--- name: support-escalation-brief description: Turn messy customer tickets, account notes, Slack excerpts, incident updates, or CRM snippets into a Support / Customer Success escalation brief. Use this whenever a support lead, CSM, or customer-facing operator needs to escalate a customer issue, summarize business impact, prepare an engineering or product handoff, identify missing context, or draft internal follow-ups for a high-priority support case. --- # Support Escalation Brief ## Intended User Role Support / Customer Success ## Workflow It Solves Convert fragmented customer issue context into a clear internal escalation brief that support, customer success, product, and engineering teams can act on without reconstructing the situation from scattered notes. ## Required Inputs - Customer issue context from tickets, chat transcripts, call notes, Slack threads, CRM notes, incident updates, or customer emails. - Account context if available: customer name, plan or tier, renewal date, ARR or revenue sensitivity, executive sponsor, SLA, open opportunities, or customer commitments. - Technical context if available: product area, environment, error messages, affected users, timestamps, reproduction steps, logs, links, prior incidents, or known workaround. - Desired audience if available: engineering, product, support leadership, customer success leadership, incident commander, or account team. If key details are missing, proceed with a labeled assumption or gap. Ask a clarifying question only when the missing detail changes urgency, customer promise, safety, legal/compliance exposure, or the recommended owner. ## Expected Outputs Produce an escalation package with: - Escalation summary: customer, issue, current state, urgency, and recommended owner. - Business impact: affected users, revenue or renewal risk, SLA/customer promise exposure, sentiment, and executive visibility when known. - Technical handoff: symptoms, environment, timeline, reproduction evidence, attempted fixes, workaround status, and links or artifacts referenced in the input. - Severity recommendation: clear severity level with rationale and confidence. - Missing context checklist: facts needed before the receiving team can act. - Recommended next actions: owner, action, deadline, and dependency. - Internal escalation note: a concise message ready to paste into an internal channel. - Customer-facing response draft when useful, clearly marked for human review. ## Procedure 1. Identify the customer, account tier, product area, escalation audience, and reporting horizon from the prompt and inputs. 2. Separate confirmed facts from inferred risk. Do not present guesses about root cause, contractual terms, or customer promises as facts. 3. Build a timeline from the first reported symptom through the latest update. Preserve important timestamps, repeated contacts, failed attempts, and customer commitments. 4. Classify impact: - Customer impact: who is affected, how many users or workflows are blocked, and whether a workaround exists. - Commercial impact: renewal, expansion, executive relationship, revenue, SLA, or public incident sensitivity. - Operational impact: support load, repeated contacts, duplicate tickets, manual work, or blocked internal teams. 5. Recommend severity from evidence: - Critical: customer production is broadly blocked, security/compliance exposure is plausible, SLA breach is active or imminent, or a strategic account has executive escalation with no workaround. - High: important workflow is blocked for a subset of users, customer deadline is at risk, workaround is weak, or repeated contacts show mounting frustration. - Medium: issue is material but contained, workaround exists, or investigation is needed before escalation. - Low: informational handoff, enhancement request, or non-urgent follow-up. 6. Select the receiving owner based on the next unblocker, not the loudest stakeholder. Use `Unassigned` when the input does not identify a likely owner. 7. Draft the internal escalation note with enough context for the receiving team to act: customer, impact, technical evidence, ask, deadline, owner, and missing context. 8. Draft a customer-facing response only if it helps the support or CS user. Keep it factual, non-defensive, and commitment-safe. Avoid promising timelines, fixes, credits, or policy exceptions unless they are already confirmed in the input. 9. End with assumptions and gaps so the human operator can clean up uncertainty before sending or routing the escalation. ## Quality Checks Before finalizing, verify that: - The severity recommendation is tied to evidence, not customer emotion alone. - The brief distinguishes customer-reported symptoms from confirmed internal diagnosis. - Every recommended next action has an owner or is marked `Unassigned`. - Any SLA, revenue, renewal, compliance, or executive-sponsor claim comes from the input or is labeled as a gap. - The internal note includes a concrete ask for the receiving team. - The customer-facing draft avoids unsupported promises and blame. - Missing context is actionable, not a generic request for "more details." ## Example Task User prompt: ```text Help me escalate this account issue to engineering. Acme Bank is on Enterprise, renewal is Aug 15, and their VP Ops is copied on the ticket. Since Friday 10:20 UTC, three admins cannot export audit logs from the compliance dashboard. The export button spins for 90 seconds and fails with "job timeout". Support tried clearing cache and a different browser. The customer has a regulator evidence deadline on Monday. We have a similar bug from March in ENG-8842 but I do not know if it is the same root cause. CSM says there is no workaround except manual screenshots. ``` Expected behavior: - Recommend high or critical severity with rationale tied to Enterprise tier, compliance deadline, executive visibility, and weak workaround. - Build a timeline with the Friday start time and Monday evidence deadline. - Separate the similar March bug from confirmed diagnosis. - Name missing technical evidence such as affected tenant, browser/network details, request IDs, logs, and export job IDs. - Draft an internal engineering escalation note with customer impact, technical symptoms, specific ask, and deadline. - Draft a cautious customer response for human review without promising an ETA.