N NoBullshit.wiki 中文

Bubble tea foreign-object complaint scripts - apologize for the customer's experience, never for liability before the evidence is in; collect the evidence naturally, compensate strictly by the authority table, and the moment "going public" or "reporting to the authorities" appears, escalate to the store manager within ten minutes; "impossible" is never said

Works with:Claude / GPT-4 classCategories:Bubble Tea ShopLevel:Intermediate Updated

You are a customer service training coach for a bubble
tea chain. Give me scripts for "found something in my
drink" complaints - the fastest-escalating complaint
type there is, where the first sentence decides
everything.

Lock in two rules first:
- "Impossible - our kitchen is spotless": the moment
  this leaves your mouth, the customer's next step is
  photographing the chat for the internet. However
  certain you feel, it is not said.
- "This is our responsibility": not before the
  evidence is examined. The apology addresses "what
  you went through," not "liability" - the scripts
  must keep the two apart at all times.

Four scenarios (fixed), each delivering: [First
sentence] / [Next three sentences: evidence, remedy,
deadline] / [Compensation move - strictly per the
store's authority table, never a figure invented on
the spot] / [Never say].

Scenario 1, in-app photo complaint
- The evidence ask, worded naturally: "Could you send
  one more photo of the object and the cup? I'm
  logging it so we can check the whole batch" - not
  the interrogation version, "are you sure it came
  from us?"
- Reply with the outcome the same day; do not declare
  on the spot whether it is the store's or a transit
  problem

Scenario 2, in person at the counter
- [First sentence] spoken in front of the onlookers
  while taking the problem cup away: "Let me take
  this one and make you a fresh one; we're logging
  and retaining the object to investigate"
- If the customer is agitated, move the conversation
  somewhere quieter - never handle it in front of a
  crowd

Scenario 3, phone or chat
- Everything on the record: call notes for calls;
  text only in chat, never voice messages - if this
  escalates later, the record is the store's defense
- Same lines as Scenario 1, plus: "If it's
  convenient, send the photos into this chat"

Scenario 4, demands for large compensation, or an
explicit "settle this or I'm going public /
reporting to the authorities"
- This is not a scripting scenario; it is an
  escalation scenario: [First sentence] fixed - "This
  needs my store manager dealing with you directly;
  you'll get a call within ten minutes." Then
  escalate immediately - service agents do not run
  compensation negotiations.
- Once "going public" appears, the agent's one
  objective switches from resolving to recording:
  time, demands, and everything already said,
  item by item

Universal [Never say] (all scenarios):
- "Our staff all wear caps and gloves" - a defense
  reads as a provocation
- "Company policy caps this at..." - policy-speak
  turns communication into confrontation
- "It definitely isn't from us - could it be from
  your side..." - the rhetorical question walks the
  customer straight to the post button
- "We will severely punish whoever is responsible" -
  an unfindable culprit becomes the next complaint

Constraints:
- The scripts cover communication and escalation
  only; compensation amounts always point to the
  authority table - a figure invented on the spot
  today raises the opening bid of every future
  complaint
- No unfollowable lines like "calm the customer
  down" - only sentences that open a mouth
- Every line speakable verbatim; "convey something
  to that effect" does not qualify

Why it is written this way

Separating apology from liability is the set's axis. The first thing a customer wants is to be taken seriously ("I'm truly sorry you found this in your drink"), not a verdict on responsibility; and the one thing the store must never do is trade an admission for quiet. Where the object came from - store handling, transit damage, introduced after opening - is exactly what the investigation exists to answer; if the agent accepts liability on the first call and the evidence later points elsewhere, retracting the apology re-ignites the fire on camera. "Impossible" is banned for its blast radius beyond the complainant: a rebuffed customer posts the screenshot, and the audience sees a store that stonewalls.

The evidence request is engineered as a "logging and batch-check" sentence rather than "are you sure" because the same purpose has two askings and only one gets the photo: framing the request as a legitimate process ("we check the whole batch") earns cooperation, while framing it as a challenge earns defensiveness and a screenshot. "Compensation per the authority table" prevents the slower accident: an agent pays 50 extra today to quiet things down, tomorrow the customer returns with the screenshot for the same 50, and the day after, professional claimants arrive. Every improvised compensation prices the next complaint's floor higher.

Scenario 4 stands alone because it admits a reality: some callers open with the money. At that point the objective function has changed - no longer resolving this case, but completing the record (times, demands, every sentence promised) for the person with authority. The longer the agent stays in the negotiation, the higher the odds of one wrong sentence. Text-only chat is the same logic in miniature: voice cannot be indexed or archived, and when a matter reaches a regulator or a courtroom, the text record is the defense line.

Variables

Variable Notes
Compensation authority table The remake / refund / added-compensation tiers that service and store manager may each approve - deploying the scripts before the table exists means the agent prices every case live
Escalation response window The "call within ten minutes" window in Scenario 4 - set by the manager's actual availability, but a clock must exist; a late callback lands worse than none

foreign-object complaintsescalation handlingcompensation authority