Set Up a WhatsApp AI Agent in 10 Minutes (No Coding, No Decision Trees)
AI agents aren't chatbots. You don't build decision trees. You tell it about your business, and it figures out how to talk to your leads. Here's how to set one up in 10 minutes.
AI agents aren't chatbots. This is the most important sentence in this article, so let's start here.
A chatbot follows a script you wrote. When a customer says something outside the script, it breaks. An AI agent understands your business and figures out how to respond. When a customer says something unexpected, it reasons about it and gives a sensible answer.
You don't "build" an AI agent the way you build a chatbot. There are no decision trees. No flowcharts. No boxes connected by arrows. You tell the AI about your business, what you offer, who you serve, the questions you want asked of new leads, and it handles conversations on its own.
That's why setup takes 10 minutes instead of 10 days.
Here's the step-by-step.
Step 1: Create Your Account and Connect WhatsApp (3 Minutes)
Go to app.replypop.com and create an account. Email, password, done. No credit card required.
Once you're in, connect your WhatsApp Business number. The process is identical to setting up WhatsApp Web: open WhatsApp on your phone, go to Linked Devices, scan the QR code on your screen.
That's the entire technical setup. No webhooks. No API keys. No Business Solution Provider contracts. No Meta Business Manager configurations.
If your customers also reach you on Instagram or Facebook Messenger, you can connect those from the same screen.
Time spent: about 3 minutes.
Step 2: Add Your Business Info and Qualifying Questions (3-5 Minutes)
This is where you teach the AI about your business. It's not programming. It's filling out forms.
What You Offer
Describe your services briefly and add rough pricing:
- Name: "Balayage"
- Description: "Hand-painted highlights for a natural, low-maintenance look"
- Starting price: "$185 for medium-length hair"
Do this for your main services. Most businesses have 5-15. If you type reasonably fast, this takes 3-5 minutes.
Who You Serve
A couple of sentences about your ideal client and the problems you solve. This helps the AI know when a lead is a good fit and when to politely redirect.
Qualifying Questions
This is the important one. What do you need to know about a new lead before your team follows up?
- What service are they looking for?
- What's their rough timeline or budget?
- Have they worked with someone like you before?
- What's their best phone number?
Pick 2-4 questions that matter for your business. The AI will ask these naturally in conversation, one or two at a time, not as a form.
Business Hours
Set when you're open. If you have different staff with different schedules, you can add that too, the AI uses it to set expectations, not to manage the actual calendar.
Time spent: 3-5 minutes.
Step 3: Turn On the AI (2 Minutes)
Here's where traditional tools would have you spend days building conversation flows. With an AI agent, there's nothing to build.
The AI already knows how to:
- Greet leads naturally
- Answer questions about your services and rough pricing
- Ask your qualifying questions
- Capture phone numbers
- Score leads and flag the hot ones to you
- Hand serious leads off to your team with full context
- Respond after hours with real information (not generic away messages)
You don't configure any of this. The AI reads the business information you entered in Step 2 and uses it to have intelligent conversations. That's the difference between an AI agent and a chatbot. A chatbot does what you told it to do. An AI agent figures out what to do.
You can optionally set a tone, "Be friendly and casual" vs "Be professional and concise", but even that's optional.
Time spent: about 2 minutes.
Step 4: Test It With a Real Message (2 Minutes)
Send a message to your own WhatsApp Business number from a personal phone. Something a customer would actually ask:
"Hey, how much is a balayage? Do you take new clients this month?"
Watch the AI respond. It'll answer the pricing question, ask a couple of qualifying questions, and aim toward capturing a phone number.
Try a few more:
- "What services do you offer?"
- "Do you have any openings next week?"
- "Do you guys do [niche service]? What's the price?"
- "Can my friend and I both come in together?"
Each message gets a natural, accurate response based on your actual business data. No "I didn't understand that." No "Please choose from the following options." Just conversation.
Time spent: 2 minutes (but you'll probably keep testing because it's satisfying to watch).
What It Handles Without You Configuring Anything
Here's the part that surprises most people. You didn't build any conversation flows, but the AI handles all of this out of the box.
Common Questions
"What are your hours?" "Where are you located?" "How much is X?" "What services do you offer?" "Do you take new clients?"
The AI answers these using the business info you provided. It understands the question regardless of phrasing. "What time do you close?" and "Are you still open?" and "How late are you open today?" all get the right answer.
After-Hours Messages
It's 10 PM. A lead messages asking about a service. Instead of an away message that says "We're closed," the AI responds:
"Hi! We open at 9 AM tomorrow. Happy to help with some quick info now. A balayage for medium-length hair starts at $185. A couple of questions so our team can reach out tomorrow with a good time..."
That's a qualified lead captured at 10 PM without any staff involvement.
Multi-Part Requests
"Can my wife and I both come in this weekend? She wants a facial and I need a haircut."
A chatbot would choke on this. The AI understands there are two people, two different services, and a shared timing constraint. It qualifies both appropriately and hands the whole thing off as one clean lead.
Follow-Up Questions
The AI remembers context within a conversation. If a customer asks about balayage pricing, then says "how long does it take?", the AI knows "it" refers to balayage. It doesn't ask "What service are you referring to?" like a chatbot would.
Returning Leads
If someone who's messaged before comes back, the AI recognizes them and picks up where the last conversation left off. No starting from scratch.
What Happens When the AI Can't Help
Not every conversation is straightforward. Sometimes a customer has a complaint. Sometimes they need something the AI can't handle, a complex custom request, a sensitive situation, a specific question about an existing appointment.
The AI recognizes these moments. It doesn't pretend to have answers it doesn't have. It lets the customer know that a team member will follow up, and it notifies you with the full conversation context.
You can also jump into any conversation at any time. The AI handles the routine. You handle the exceptions.
The "But What About..." Questions
"Does this replace my booking software?"
No, and that's a good thing. Replypop works alongside whatever booking tool you already use, Acuity, Calendly, Fresha, Booksy, Boulevard, Vagaro, Google Calendar, pen and paper. It handles the messaging and lead qualification. Your booking software still owns your actual schedule.
"What if the AI says something wrong?"
The AI only knows what you told it. It won't make up prices or invent services. If a customer asks about something you haven't added to your business profile, the AI says it doesn't have that information rather than guessing.
"What about WhatsApp's messaging rules?"
WhatsApp has a 24-hour messaging window and rules about template messages. Replypop handles all of that.
"Can I still message customers myself?"
Absolutely. The AI handles conversations, but you see everything in real time. Jump in whenever you want. The AI steps back when you take over.
"What does it cost?"
The free plan gives you 5 monthly credits so you can test Replypop on real inbound leads.
For the full AI agent experience, natural conversations, lead qualification, and lead-intent alerts, the Pro plan is $29/month with a larger monthly pool of AI-powered lead conversations.
Most small businesses either start on Free to validate the flow, or move to Pro once inbound volume justifies full-time coverage.
Before and After
Before: Lead messages at 9 PM. Gets an away message. You respond at 8 AM. Lead has already cooled off or found someone else.
After: Lead messages at 9 PM. The AI sales agent answers their questions, qualifies them, captures phone number. You wake up to a high-intent lead in your inbox.
Before: Lead asks a question your chatbot wasn't programmed for. Gets "I didn't understand that. Please choose from the menu." Lead leaves.
After: Lead asks a question the AI wasn't specifically prepared for. AI reasons about it based on your business information and gives a helpful answer.
Before: Saturday morning. 15 messages asking about services and availability. Your receptionist spends 2 hours responding one by one.
After: Saturday morning. 15 messages. AI handled all of them overnight. 8 qualified leads in your inbox with full context. Your receptionist focuses on the ones who need a real follow-up call.
10 Minutes. That's It.
The entire thing, account, WhatsApp connection, business info, live AI, takes 10 minutes.
Not 10 minutes of coding. Not 10 minutes of building decision trees. Not 10 minutes of the easy part before 10 hours of the hard part. Actual, total, 10 minutes.
The difference between an AI agent and a chatbot is the difference between hiring someone smart and writing an instruction manual. One takes 10 minutes of explanation. The other takes weeks of documentation and still doesn't handle surprises.
Your WhatsApp number is already getting messages. The only question is whether those messages turn into qualified leads or away messages.
Questions or feedback? Reach out anytime