Try this skill in Strawberry

/escalate-customer-issue

Escalate Customer Issue

Give the receiving team enough grounded context to act without making them reconstruct the customer history. Preserve Support or Customer Success ownership of the customer communication loop.

1. Confirm the need and route

Establish the support case or source link, customer-facing owner, current communication state, impact, timeline, prior attempts, workaround, and why the normal path is insufficient. Choose the specialist from the actual issue rather than defaulting technical work to one project or owner.

Use breadth, depth, duration, workaround, contractual obligation, and accepted relationship context to apply the team's priority definitions. Do not inflate priority because an issue is technical. Follow the accepted urgent path for security, privacy, data loss, legal, regulated, or active incident cases.

2. Build the evidence and ask

Use the complete customer conversation plus relevant account context, product evidence, logs, screenshots, team discussion, and linked work. Preserve source links, timestamps, versions, environments, prior attempts, and unknowns. Keep customer data to the minimum the destination needs.

State the exact ask: investigate, reproduce, decide, approve an exception, provide a safe response, or commit an owner and update time. Distinguish the internal escalation, engineering issue creation, and customer response as separate actions.

For a product failure, use strawberry/product-engineering/report-bug to research, deduplicate, prepare, file, or update the engineering issue. This skill owns the impact, relationship context, handoff, and communication obligation.

3. Deliver and record the handoff

Provide:

  • one-line summary, case or source link, customer-facing owner, and target team;
  • customer impact and calibrated priority with evidence;
  • timeline, previous attempts, workaround, and current state;
  • technical or product evidence and linked work where relevant;
  • last customer update and current communication state;
  • exact ask, decision deadline, open questions, and next update; and
  • brief customer acknowledgement copy for use after the handoff occurs.

Posting internally, assigning an owner, filing or updating an issue, and replying to the customer are separate external actions. Follow current scoped authorization and verify each completed action.

After the handoff, record the linked owner, named dependency, and next update in the case record. Send the acknowledgement only after the escalation has happened. Keep the customer-facing owner responsible until the outcome is verified and the customer loop is complete.

A customer escalation often fails in one of two directions. It is either too thin for the receiving team to act on, or it becomes a long ticket history with no clear question. Meanwhile, the customer is still waiting for an update.

Strawberry can bring together the ticket, account history, team discussion, product evidence, previous attempts, and current commitments in one visible workspace. Your companion can prepare the internal brief and the customer update from the same evidence while keeping those audiences separate.

Explain why the normal path is not enough

Start with what is happening, who is affected, how long it has persisted, what has been tried, and what changed now. Use the team's escalation and incident policy to choose the route. Suspected security, privacy, data loss, legal, regulated, or active incident cases should follow the urgent specialist path immediately.

After a few useful handoffs, save the brief structure, routes, and ownership rules as a team skill. Keep each customer's history and sensitive details in the case they belong to rather than baking them into the shared method.

Build one evidence-backed timeline

  • Link the original request, material customer updates, and attempted resolutions.
  • Preserve timestamps, affected environments, exact errors, screenshots, and relevant logs.
  • Separate verified impact from inferred churn, revenue, or reputation risk.
  • Show the current workaround, owner, and next promised customer update.

Connected context reduces reconstruction, but it should not create a data dump. Include only the commercial, account, and technical detail the receiving team needs, and keep private or sensitive context out of broadly visible records.

Split customer impact from product evidence

OwnerWhat it contributes
Customer Support or SuccessCustomer impact, relationship history, escalation timing, current communication, and next update.
Product and EngineeringReproduction, environment, technical context, duplicate search, and the engineering issue.
Receiving specialistThe investigation, decision, approval, or exception the escalation requests.

Ask for a specific next move

State whether the team needs an investigation, reproduction, decision, safe response, approved exception, owner, or update time. Then track that commitment and maintain the customer communication until the outcome is verified. Posting the escalation, filing an issue, sending the update, and offering a commercial remedy are separate actions.

/close-open-loops

Show me what's slipping through the cracks.

View skill
How to research and report software bugs with AI

Carry a product problem into reproduction, duplicate search, and a grounded engineering issue.

How to track commitments with AI

Keep cross-team owners, decisions, and promised updates from slipping through the cracks.

How to write better status updates with AI

Turn the reviewed escalation picture into a concise update for the people who need it.