Reframe a Public Dressing-Down From a Client Into a Structured Debrief for Your Team

A client tore into your delivery on a call, in front of colleagues, and now you have to walk into the team room and say something useful rather than relay the sting. This separates the fair, checkable criticism from the noise, gives you three fixes you can start this week, and hands you the line to open the debrief with.

<context>
You are an experienced account lead who has just taken a hard verbal knock from a client on a call, in front of colleagues, and now has to walk back into the team room and turn it into something useful rather than pass the sting along. Your task is not to relitigate whether the criticism was fair. It is to convert what was said into a structured account the team can act on.
</context>

<paste>
{WHAT_WAS_SAID}
</paste>

<task>
Work only from the content provided in {WHAT_WAS_SAID}. If a section below cannot be answered from that content, state that explicitly rather than inferring what the client meant.

1. Separate the criticism into two lists: specific, checkable complaints, and general venting or tone that carries no new information
2. For each specific complaint, state what evidence would confirm or dispute it
3. Draft three fixes the team could realistically start this week, tied to the specific complaints only
4. Draft the one sentence you will use to open the team debrief, calibrated to inform without alarming
</task>

<output_format>
- **Specific complaints**: bulleted, one line each, with the evidence needed to check it
- **Noise**: bulleted, the venting or tone that is not actionable, kept separate so it is not accidentally treated as a task
- **Three fixes**: numbered, each one sentence, realistic within a week
- **Opening line for the team**: one sentence, factual, no blame language
- Length: roughly 250 to 350 words total
- Tone: composed and factual throughout. Do not let anger or defensiveness into any section
- Edge case: if {WHAT_WAS_SAID} is a rough or partial recollection rather than a transcript, say so at the top of the output and mark any inferred detail as inferred rather than quoted
</output_format>