How to Build an Instagram Lead Qualification AI Agent
An Instagram DM can start with a very simple message:
Hi, I'm interested.
But your sales team needs much more information.
They may need to know:
Who the person is
What they need
Their budget
When they want to start
Whether they are actually ready to buy
Instead of asking your team to manually qualify every Instagram enquiry, you can build an AI Agent that handles the first part of the conversation.
What the agent should do
A good lead qualification agent should:
Understand the enquiry.
Answer basic questions.
Ask useful qualifying questions.
Save the answers.
Classify the lead.
Send qualified leads to sales.
It should not interrogate every customer immediately.
The conversation should still feel natural.
Example conversation
Customer:
Hi, I'm interested in your service.
AI:
Absolutely. I can help with that. What are you looking to achieve?
Customer:
I need WhatsApp automation for my business.
AI:
Great. What's the approximate monthly message volume you're expecting?
Customer:
Around 20,000.
AI:
Thanks. And what's your name?
Customer:
Rahul.
The AI can continue until it has the information required for qualification.
Decide your qualification questions
Don't blindly copy somebody else's questions.
Choose the information your sales team actually uses.
For example:
Name
Requirement
Budget
Timeline
Maybe:
Company size
or:
Location
Create the agent
Go to the AI Agents area.
Create:
Instagram Lead Qualification Agent
Add a description explaining the job.
Write the System Prompt
A useful starting prompt is:
You are the Lead Qualification AI Agent for our Instagram account.
Your job is to help new Instagram enquiries and identify serious prospects.
Be friendly, conversational and concise.
First understand what the customer wants.
When appropriate, collect:
- Name
- Email
- Requirement
- Budget
- Timeline
Do not ask all questions at once.
Ask naturally based on the conversation.
Use Knowledge to answer questions about our business.
Never invent pricing, availability or company information.
Once the required lead information has been collected:
1. Save the customer information.
2. Add the appropriate lead label.
3. Hand the conversation to the sales process.
If the customer asks to speak to a human, hand the conversation to the sales team.
Add Knowledge
Give the agent the information it needs to answer basic enquiries.
For example:
What does your company do?
Which industries do you serve?
What services do you offer?
What are your business hours?
What are your pricing ranges?
The agent should use your configured Knowledge rather than guessing.
Add Save Custom Field actions
Suppose you have custom fields for:
Name
Email
Requirement
Budget
Your agent can use:
##save_custom_field## : ID
for the appropriate fields.
The action IDs shown in ZazzyAgent should be selected using the action autocomplete rather than guessed manually.
Add a lead label
Once the customer has provided enough information, the agent can add an appropriate label.
For example:
##add_label## : ID
You might use:
Qualified Lead
or:
Instagram Lead
depending on your CRM structure.
Send the lead to sales
Once qualification is complete, the conversation can be transferred to another AI Agent or a human.
For example:
##transfer_agent## : ID
or:
##assign_human_role## : ID
The choice depends on whether you want another AI Agent or a human sales representative to take over.
Don't qualify every conversation
This is important.
Someone saying:
Nice post!
isn't necessarily a lead.
Someone saying:
I'm looking for a service for my company. Can you tell me the price?
probably deserves a useful response.
The AI should understand the difference.
Don't ask questions in a rigid sequence
A poor bot says:
What is your name?
What is your email?
What is your budget?
What is your requirement?
even when the customer has already answered some of these.
A better agent watches the conversation and only asks for missing information.
Example
Customer:
Hi, I'm Rahul. I run a real estate company and need an automated lead system. Budget is around ₹1 lakh.
The agent already knows:
Name
Industry
Requirement
Budget
It shouldn't ask:
What's your name?
again.
It should ask for only what is missing.
Define what "qualified" means
Your System Prompt should explain when a lead becomes qualified.
For example:
Consider a lead qualified once name, requirement, budget and timeline have been collected.
You may use a different definition.
The important part is that the AI has a clear rule.
High-intent lead example
Customer:
I want to start next week. Can someone call me today?
This is a strong signal.
Your agent should not continue asking unnecessary questions.
It can collect anything essential and then hand the conversation to sales.
Human handoff
Give customers a clear path to a person.
For example:
I'd like to talk to someone.
The AI should stop trying to automate everything.
It can assign the conversation to your sales team.
Test your lead agent
Test at least:
Curious visitor
Just checking what you offer.
Pricing enquiry
How much does it cost?
Strong lead
I need this next week.
Incomplete lead
I'm interested.
Human request
Can someone call me?
Unrelated message
Good morning!
Your agent should behave appropriately in all six situations.
Track the information
After testing, check the customer record.
Confirm:
Custom fields were saved.
Label was added.
Correct team/agent received the conversation.
Customer did not have to repeat information.
Recommended architecture
Instagram DM
↓
Lead Qualification Agent
↓
Understand Intent
↓
Answer Basic Questions
↓
Collect Missing Information
↓
Save Custom Fields
↓
Add Lead Label
↓
Qualified?
↙ ↘
No Yes
↓ ↓
Continue Sales Agent
or Human
A lead qualification agent works best when it feels like a helpful sales assistant—not a form disguised as a chatbot.