Automate replies across TikTok and YouTube with an AI engine that reads comments, writes responses and captures leads.
A fully automated comment management system that pulls comments from TikTok and YouTube, understands sentiment and intent, generates natural human level replies and pushes high intent commenters into your CRM. Users maintain constant engagement, handle every comment at scale, convert hidden leads and grow their audience without manually checking notifications or typing responses.
Set up an Attio workspace with a pipeline or list called Social Engagement and create fields for platform, video URL, commenter handle, comment text, sentiment, intent, reply_sent, lead_score and status.
In n8n, create a workflow that regularly pulls recent comments from your TikTok and YouTube content using their APIs or a connected comment export service and passes each new comment into the workflow with metadata like video ID and author.
Add an n8n node that sends each raw comment and basic context, such as platform, video topic and caption, into Relevance AI so it can classify the comment as positive, neutral, negative, spam, question or high intent and extract key phrases or topics.
Configure Relevance AI to return a structured JSON object for each comment that includes sentiment, intent label, spam flag, topic tags and an initial lead_score suggestion.
Pass the original comment, the video context and the Relevance AI labels into Claude via n8n and ask it to generate a short, human sounding reply that matches your brand tone and respects the intent and sentiment of the commenter.
In the same Claude call, request that it also output a short internal note for you, such as whether to invite the user to DM, link to an offer, or simply acknowledge and encourage more engagement.
Map Claude’s outputs into fields for public_reply and internal_note and combine them with the Relevance AI fields so the full enriched comment record is ready for storage.
Use an n8n Attio node to create or update a record in your Social Engagement list for each commenter, storing the comment text, sentiment, intent, public_reply, internal_note, lead_score and status.
Add a conditional step in n8n so that if Relevance AI or Claude flags a comment as spam or low value, the workflow stops there and marks the comment as Ignored in Attio without posting a reply.
For all valid comments, add a node in n8n that posts Claude’s public_reply back to TikTok or YouTube via their respective APIs so the answer appears under the original comment, then set reply_sent to true in Attio.
Create an additional branch in n8n that, for comments marked as high intent or strong buying signals, updates the Attio record status to Lead and optionally sends you a notification so you can follow up manually with a DM or personalised offer.
Schedule a separate daily n8n workflow that queries Attio for all high intent and high lead_score commenters, sends a summary into Claude to identify patterns or new content ideas and writes the insights back into Attio so you can adjust your content and reply style over time.
Build and deploy custom AI agents to automate business workflows.
Meet your AI thinking partner for any complex task.
CRM re-engineered for high-velocity revenue teams.
Build custom workflows with code-level power and no-code speed.
This blueprint builds an autonomous comment reply system that monitors your TikTok and YouTube videos, understands what people are saying, writes human sounding responses and pushes high intent comments into your CRM so you never miss a potential lead. Instead of manually scrolling through notifications and typing replies all day, the workflow runs continuously and handles everything from sentiment analysis to personalised responses.
The system begins with Attio, which acts as the memory layer. You create a simple pipeline that stores every comment and its metadata, including the platform it came from, the comment text, sentiment, intent, suggested reply and whether a response has already been posted. This lets you treat comments the same way a salesperson treats inbound leads. Every interaction is tracked, every high value comment is surfaced and every potential customer is visible.
The automation engine is n8n. This is where the full loop runs. n8n regularly pulls new comments from your TikTok and YouTube accounts. It collects the comment text, the username, the video it belongs to and any context available, such as the caption or video topic. Every time a new comment appears, n8n triggers the rest of the workflow.
Relevance AI acts as the classification layer. Each raw comment is sent into Relevance AI so it can detect sentiment, intention, spam, praise, questions and buying signals. Relevance AI also extracts key phrases or themes such as whether the commenter is expressing confusion, interest, positivity or frustration. This gives you clean structured data instead of raw comments that you need to interpret yourself.
The enriched comment is then passed into Claude. Claude handles the reply generation. Using the original comment, the sentiment and intent labels and the video context, Claude writes a short, natural, on brand reply that feels human and conversational. Claude can adjust tone, match your persona and handle a wide range of interactions, from positive encouragement to polite corrections. In the same prompt, Claude also generates an internal note that tells you whether you should follow up manually, invite the commenter to DM or leave it as simple engagement.
The workflow now has everything it needs. n8n maps the response, sentiment, intent and internal note into Attio and creates or updates the commenter’s record. You now have a database of people who interact with your content, organised by intention and potential value.
Before posting any reply, the workflow checks whether the comment is spam or meaningless. If the comment is low value, n8n marks it as ignored in Attio and stops. This prevents your pages from being cluttered with unnecessary interactions.
For valid comments, n8n posts Claude’s reply directly under the comment on TikTok or YouTube using their APIs. Once the reply is posted, the workflow updates Attio so reply_sent is marked true. You now have a clean record of what was replied to and what was not.
High intent comments get special treatment. If Relevance AI or Claude detects strong buying signals, the workflow moves that commenter into a Lead status inside Attio and optionally notifies you so you can follow up with a DM. This turns YouTube and TikTok comments into a real sales pipeline.
Finally, to make the system smarter over time, n8n runs a daily job that pulls all high intent comments from Attio and sends them into Claude to generate insights. Claude identifies patterns in what people are responding to, what objections are recurring and what topics are generating the strongest audience engagement. The insights are stored back inside Attio so you can refine your content strategy and reply tone.
The result is a self sustaining loop that monitors your content, interprets comments, replies automatically, captures leads and produces feedback that improves your content over time. You stay active across TikTok and YouTube without needing to manually type responses, and you never miss a high intent comment that could turn into a customer.
Share it with your network