Draft real, on-brand replies to actual complaints Grok found on X

You are finding specific, current customer complaints about a brand on X using Grok's live post access, and drafting a tailored reply option for each one — not a single generic apology template that gets pasted under every complaint regardless of what it actually says. A complaint about a billing error and a complaint about a rude support interaction need genuinely different replies, not the same we're-sorry-to-hear-this line with the specifics swapped out.

BRAND
A mid-size meal-kit delivery service

TIME WINDOW
Past 24 hours

BRAND VOICE AND ESCALATION POLICY
Warm, first-name, no corporate jargon; escalation policy says any complaint mentioning a missing or spoiled delivery must move to DM or a private channel, never resolved with a public promise of a specific refund amount

WHAT THE REPLY CANNOT DO
Never promise a specific refund amount or timeline in a public reply; never confirm or deny an operational detail, like which supplier caused an issue, that hasn't been internally confirmed

PRIORITIZATION CRITERIA
Complaints from accounts with visible follower counts over 5,000 first, then complaints mentioning food safety specifically, then everything else by how long it’s gone unanswered

REPLY RULES
Find complaints specifically, not general mentions — a complaint has a clear grievance and often an implicit or explicit ask, while a neutral or mixed mention does not need a reply drafted for it at all, so filter for the former before drafting anything. Read each complaint's specific content closely enough that the drafted reply responds to what was actually said, not to the general category of complaint it belongs to — if the complaint names a specific order number, a specific interaction, or a specific broken feature, the reply should reference that specific detail, not a generic acknowledgment that could apply to any complaint of that type. Match the reply's tone and length to the platform and the specific complaint's intensity — a brief, calm complaint gets a brief, calm reply; a complaint written in visible frustration gets a reply that actually acknowledges the frustration in its opening line rather than jumping straight to a procedural next step, since a purely procedural reply to visible anger reads as tone-deaf regardless of how correct the procedure itself is. Never draft a reply that makes a commitment the brand voice and escalation policy doesn't actually authorize — a refund promise, a guarantee about resolution timing, an admission of fault on a matter that's still being investigated — and where the honest answer requires escalating to a private channel rather than resolving publicly, the drafted reply should say that plainly rather than pretending the public reply itself resolves anything. Respect every item in the reply constraints list as an absolute boundary, not a style preference to be weighed against sounding maximally reassuring in the moment. Prioritize which complaints get a reply drafted first using the stated criteria — visibility, severity, how long it's been unanswered — rather than working through them in whatever order they happened to surface in the search, since a support or comms team acting on this output needs the most urgent ones addressed first, not the easiest ones to write a reply for.

OUTPUT FORMAT
1. A prioritized list of complaints found, each with the account (redacted if not public or verified), the complaint quoted, and the priority reason.
2. For each, a drafted reply, and a one-line note on why that specific tone and length were chosen for this complaint.
3. Any complaint you deliberately did not draft a public reply for because it needs private escalation instead, with a one-line reason.
4. Anything that came up in the search that didn't qualify as a complaint and was correctly excluded.