As of 13 August 2026, AI can reply to comments on your social media posts.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 15 minutes until you can act on the result.

   How to actually do it

   1. Open the social platform and copy the comments you want to answer into a document, keeping each comment separate and including the post it refers to.
   2. Gather your current brand voice, approved product facts, response policy, escalation contacts and any current offers, then remove customer names, order details and other private information unless it is essential and authorised.
   3. Paste the comments and the gathered guidance into the prompt, setting the reply word limit and stating whether emojis, links or direct messages are allowed.
   4. Ask the chatbot to produce the numbered reply table and to mark complaints, account-specific requests, threats and uncertain cases as ESCALATE.
   5. Compare every drafted reply with the original comment, your current facts and your response policy, then rewrite any reply that sounds defensive, makes an unsupported promise or exposes private information.
   6. Paste only the approved replies into the relevant social posts, send escalated cases through your agreed private or internal route, and record any new response rule your team needs.

   Prompt

   You are helping manage social media for [business or organisation]. Draft replies to the comments below.
   
   Use this brand voice: [plain description of tone].
   Use these approved facts and offers only: [paste facts, product details and current offers].
   Follow these rules: do not invent facts, make promises, discuss private customer information, argue with commenters or give legal, medical or financial advice. Flag complaints, threats, safeguarding concerns, requests for refunds, press enquiries and anything requiring account-specific information as ESCALATE instead of drafting a public reply.
   
   For each comment, provide:
   1. The original comment.
   2. A suggested public reply of no more than [number] words.
   3. The purpose of the reply.
   4. Any fact or decision I must check before posting.
   5. ESCALATE where the comment should be handled privately or by a named colleague.
   
   Keep replies specific to the comment, natural rather than promotional, and do not use emojis unless the comment already uses them. Do not claim that you have checked the live post or customer account.
   
   Comments:
   [paste comments here]

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.