Inrō’s AI Agent turns your Instagram DMs into a smart assistant that can qualify leads, answer questions, and trigger your best automations.
The AI Agent is no longer a separate add-on – it’s built into Inrō, and it’s part of the Pro Plan starting at €12.99. Once your Instagram is connected to Inrō, you can open the AI Agent section and walk through a simple step-by-step setup.
💡 Unlike other automation tools on the market, Inrō's AI agent is included with the Pro plan and has no usage limits.
In this article, you will learn how to:
Define your Inrō AI agent’s persona
Give it clear goals
Control how much it can answer based on your knowledge
Configure human handoff
Map user intentions to concrete actions
Enrich it with documents
Use the right-hand side text window, which you’ll always see, to test and try the agent as you configure it.
Step 1 – Define your AI persona
Your persona tells the AI who it is and how it should talk.
In the Persona step, you can:
Detect from your profile
Let Inrō scan your connected Instagram account to auto-generate a first version of your persona (who you are, what you sell, how you speak). This is perfect for getting started quickly.
Write your own persona
You can also write or edit everything manually:
Who you are with the role and behaviour
How the agent should speak (tone, formality, emojis, languages, etc.
Step 2 – Give your agent a goal
Next, you’ll tell the AI what success looks like.
In the Goals section, you’ll see:
Account type
Choose what best describes you (creator, coach, e-commerce, local business, etc.). This helps us pre-configure your agent for your use case.
Account description
Explain what your account does, what makes it different, and the value you bring to followers.
Primary goal
What is the main outcome you want from the AI Agent?
Examples: “Book free calls with qualified leads,” “Sell my online course,” “Drive people to visit my bar.”
Keep this clear and action-oriented.
Other goals
Any additional outcomes you’d like the agent to support (e.g. grow email list, send resources, qualify leads). This gives the AI more context about your priorities.
Key proposition
A short sentence that answers: “Why should someone choose me?”
The agent will reuse this line to highlight your strengths in sales or support conversations.
Audience type
Describe your ideal followers (who they are, what they need, their level of knowledge). The AI will adapt its replies to this audience.
Additional instructions
Any extra rules about how your AI should behave: what to always do, what to avoid, how pushy or soft it should be, etc.
This keeps the agent aligned with your brand voice and boundaries.
💡 Tip: Ask yourself: “If my AI could only do one thing perfectly, what would it be?” That’s your primary goal.
Step 3 – Control knowledge access
In the Knowledge Access section, you can decide how strictly your AI Agent should rely on your own data.
You have two main options:
1. Answer only when knowledge exists
Tell the AI to reply only if it finds an answer in your knowledge.
If it can’t answer:
The conversation is routed to a dedicated folder for your team to handle manually.
Your team receives an email notification.
The user sees a default message like:
“Sorry, I don’t have an answer to this question. Someone from the team will be in touch soon.”
This keeps the AI 100% on-brand and prevents it from guessing or inventing details.
2. Always have an answer
Let the AI Agent use both:
Your provided knowledge, plus
Its built-in language model understanding
In this mode, the agent will always try to respond, even if your documents don’t cover the question.
Step 4 – Allow human intervention
Your AI Agent knows when to step back and let a human take over.
In the Human intervention / Handoff settings, you can:
Decide that the AI should stop replying so your team can answer from the Inbox
Optionally tag or flag conversations that need human review
Define how the agent should respond when someone asks for a human
This keeps conversations feeling safe, human, and under your control.
Step 5 – Define intent-based goals & actions
Now you’ll connect user intentions to specific actions the agent can take.
In the Goals & actions step you can:
Create goals based on user intent
For example:
“Book a free call”
“Get a course link”
“Join my email list”
“Get a quote”
For each goal, choose what the AI should do when it detects that intent:
Send conversion links
Share a specific checkout link, booking link, or landing page when the person is ready.
Activate a Scenario
Trigger one of your existing Inrō automations (e.g. a full call-booking flow or coupon flow).
Add people to a Folder
Segment users into folders (e.g. “Warm leads”, “Course interested”, “VIPs”) for future campaigns.
The agent will read the conversation, detect the user’s intention, and then trigger the right action behind the scenes.
💡 Tip: Start with 1–2 key goals (e.g. “Book calls” and “Sell my main offer”) and add more later.
Step 6 – Import documents
Finally, you can enrich your AI with supporting documents.
In the Documents / Knowledge area, you can:
Import PDFs, guides, FAQs, menu/price lists, or course overviews
Paste existing documentation from Notion, Google Docs, etc.
Keep everything organized by topic so it’s easy to update
The AI will use these documents to answer questions more precisely (e.g. “What’s included?”, “How much is it?”, “What are your opening hours?”).
Testing your AI Agent
Throughout all these steps, you’ll see a live test chat on the right.
You can:
Ask the AI the same questions your followers usually ask
Check that tone, persona, and goals feel right
Test specific flows (e.g. “I want to book a call”, “Do you have a size guide?”)
Make sure it triggers the right scenarios, folders, or links
Adjust your persona, goals, knowledge access or actions, then test again until it feels natural.
Going live
Once you’re happy with how the AI behaves in the test chat:
Save / Update your AI Agent settings.
Activate it at the top of the page
Integrate it into your live flows, for example, by:
Enabling AI inside specific Scenarios
Letting the agent handle certain types of incoming DMs by default
Handing over only specific flows to the AI agent.
To know more about activating the AI in specific scenarios, read this article.
💡 Pro Tip: If you prefer the AI Agent to trigger only in specific automations, leave the global activation switch off and add the AI Agent step in your custom scenarios.
Feedback
If you have any questions or ideas to improve the AI Agent experience, we’d love to hear them. Your feedback helps us make the agent smarter, safer, and more effective for your business.