If your AI Agent isn't sending replies, one of the settings below is usually why. Work through the checks in order.
Check the activation mode
This is the most common cause. The agent has four activation modes, and each one changes when replies are allowed. Go to AI Agent → Activation rules and see which is selected:
All messages: The agent replies to every message that an automation hasn't already handled. If this is on and replies still aren't going out, keep reading.
Specific intents only: The agent only replies when a message matches one of your configured intents. If no intents are set up, the agent replies to nothing. Add intents, or switch to All messages.
Automation only: The agent ignores normal inbound DMs and only replies when a scenario hands a conversation to it. If you expect it to answer DMs directly, this mode is why it stays quiet. Switch to All messages or Specific intents only.
Disable all replies: A full off switch. The agent won't respond to anything, even when a scenario hands off to it.
For the full breakdown of each mode, see Activation rules: controlling when your AI Agent replies.
Have your AI credits run out?
If the agent was working and went silent on its own, check your AI credits. This catches a lot of people by surprise.
On the free plan, the AI Agent runs on a pool of free credits. Every reply and AI action uses one. When the pool hits zero, Inrō automatically switches the agent to Disable all replies as a safety stop, so it stops responding even though you didn't change the setting yourself.
To get the agent running again, upgrade to Pro, which includes unlimited AI. Then set the activation mode back to the one you want. See AI Credits: what they are and how to check them to view your balance.
Check folder rules
Activation rules can include folder restrictions. If a contact is in a folder that's excluded from the agent, the agent won't reply to them. In the activation settings, look for any folder rules and confirm the contact isn't being excluded.
Is a scenario running for that contact?
Scenarios take priority over the AI Agent. If a contact is mid-flow in an active scenario, the agent won't interrupt it.
Open the contact's conversation in the Inbox and look at the handling status next to their name. If it shows the yellow lightning bolt (Automation), a scenario is running, and the agent will take over once the scenario finishes or hands off. See Handling status: human, automation, agent, and more for what each icon means.
Is the conversation in human handling mode?
If a team member replied manually, Inrō may have switched the conversation to Human handling, and the agent steps back. Check the handling status in the conversation header. If it shows the Human icon, switch it back to Agent from the header when you're ready. See Human intervention: handing off to a human.
Check your Instagram connection
If the link between Inrō and Instagram is broken, the agent can't send anything. Go to Accounts → Instagram account and confirm it shows Connected. If it's disconnected, see How to fix common Meta & Instagram connection errors.
Is it the knowledge base?
An empty knowledge base does not stop the agent from replying. The agent still answers from the instructions and goals you've set. So if replies aren't coming, the knowledge base isn't the cause.
🐾 Netsuke's Tips
Start with the activation mode. The two we see most often are "Disable all replies" and "Specific intents only" with no intents configured. Both leave the agent silent.
If the agent stopped on its own and nothing in the settings changed, check your AI credits before anything else. The free-credit safety stop is the usual culprit.
An active scenario always wins over the agent. If you want the agent to take over mid-flow, add a "Hand over to AI Agent" action inside your scenario to pass control on purpose.
⚠️ If none of these fix it
Reach out to our support team with a description of what's happening, which contact the agent isn't responding to, and a screenshot of your activation settings. We'll get it sorted.