Your Instagram Auto-Reply Is Legally a Chatbot Now. Here's the One-Line Fix

12 min read By FindSkill Team

From August 2, EU rules say customers must know when they're talking to AI. If your shop runs auto-replies on Instagram or WhatsApp, one greeting line fixes it.

Table of Contents

More than a million businesses were already using Meta’s AI Business Agent at its June launch — the assistant that answers “is this still available?” and “do you ship to Berlin?” on WhatsApp and Instagram while you sleep. Add the shops using ManyChat, Chatfuel, or a website chat widget with AI behind it, and a huge share of small businesses are already running a chatbot, whether they’d use that word or not.

Here’s the thing that changes on August 2, 2026: under Article 50 of the EU AI Act, customers interacting with an AI system have the right to know it’s AI. If EU customers can message your shop and a machine answers, that’s now your obligation — you, the business deploying the bot, not just Meta or the chatbot company that built it.

The fix costs you one sentence. Let’s do it properly.

What the rule actually says

Article 50(1) of the EU AI Act requires that people interacting directly with an AI system are informed they’re dealing with AI — unless it’s already obvious to a reasonably attentive person from the context.

Three things worth understanding before you touch your settings:

Provider duty
Meta, OpenAI, and chatbot platforms must build systems so users can tell they're AI — design-level disclosure and machine-readable marking of content.
Deployer duty (this is you)
If your business runs an AI assistant that talks to customers, you make sure the customer knows. In practice: the bot introduces itself as AI.
the tool maker who owes what you, the shop
  • “Obvious from context” is a real exemption, but don’t lean on it. A robot avatar named “AI Assistant” is arguably obvious. An auto-reply that just says “Hi! How can I help you today? 😊” from your shop’s name and profile photo is not. Ambiguity is exactly what the disclosure line removes.
  • This applies to EU customers, wherever you are. Like GDPR before it, the AI Act follows the user, not the seller. If EU buyers can DM your shop, the clean move is to turn the disclosure on for everyone rather than geo-splitting your greeting.
  • The fines are corporate-scale but real — up to €15 million or 3% of worldwide turnover for transparency violations. Nobody expects a market-surveillance authority to open with a craft shop; the point is that compliance costs one sentence and non-compliance is now a legal exposure, not just a style choice.

Does your setup even count as a chatbot?

Thirty-second check: do you owe the disclosure line?
Customers can message your shop? Instagram DM, WhatsApp, website chat
Something answers automatically? no → you're a human, no label needed
Does it converse? static 'we reply within 24h' text → exempt
AI answers = add the line Meta Business Agent, ManyChat AI, GPT widgets
Only conversational AI triggers the duty — static auto-replies and human typing don't

The one-line fix

Open your auto-reply or chatbot greeting and make the first message say what it is. Any of these work:

Hi! I'm the AI assistant for [Shop Name]. I can answer questions
about products, shipping, and orders. If you'd like a human,
type "person" and [Name] will reply during shop hours.

Thanks for messaging [Shop Name]! You're chatting with our AI
helper — I'll do my best, and anything I can't handle goes
straight to a real person.

The pattern: name it as AI in the first message, and say how to reach a human. The second half isn’t legally required by Article 50, but it’s what keeps the disclosure from feeling like a downgrade — customers accept bots happily when the escape hatch is visible.

Where to set it:

  • Meta Business Agent / Instagram auto-replies: Business Suite → Inbox → Automations — edit the greeting and FAQ responses so the first line includes the AI mention.
  • ManyChat / Chatfuel / website widgets: edit the welcome message block in your flow. One block, done.
  • A human-written auto-reply (“We’ll get back to you within 24 hours”) is not a chatbot — no disclosure needed for a static text message. The rule is about AI that converses.

While you’re in there, check what the bot is allowed to answer. The disclosure line fixes the legal question; it doesn’t fix a bot that confidently invents your return policy. Review its answers to your top five customer questions the same way you’d proofread an employee’s saved replies.

What this means for you

If you already use Meta’s Business Agent: Meta has been rolling out AI-identification into these products, but the deployer duty is yours — open your greeting today and confirm the AI mention is actually there in the first message your customer sees. Thirty seconds to check beats assuming.

If you sell to the EU from outside it: same as your privacy policy — the practical answer is one global setting, not a border check. The disclosure line doesn’t hurt you anywhere else; US customers increasingly expect it too.

If your “chatbot” is just you, typing fast: you’re fine. Humans don’t need a disclosure label. Yet.

If you’re also labeling AI product photos and listings: that’s the other half of the same August 2 rule — we covered exactly what Etsy and the EU require for AI-made content, including the viral claims that turned out to be false.

The bottom line

From August 2, an AI that talks to your customers has to introduce itself as one. One greeting line satisfies the rule, keeps trust when customers figure it out anyway (they do), and takes less time than reading this post did. Set it, note the date you did, and get back to selling.

If you want your AI assistant to actually be good — trained on your policies, escalating the right conversations, drafting replies in your voice — our Customer Service Excellence with AI course covers the full setup, and Claude for Small Business shows the wider workflow.

Sources

Build Real AI Skills

Step-by-step courses with quizzes and certificates for your resume