WhatsApp + OpenAI: brochure based replies that sell
You know the feeling. A customer asks a simple product question on WhatsApp, and suddenly you are digging through a PDF brochure, scrolling, guessing, then rewriting the same answer you sent yesterday.
Sales reps get stuck in “quick questions” all day. A marketing manager running a launch feels it too, because every delayed reply costs momentum. Even founders end up doing support-by-WhatsApp. This WhatsApp OpenAI replies automation gives you fast, brochure-accurate answers without living inside your catalog.
In a few minutes, you will understand what the workflow does, what you need to run it, and how to customize it so it sounds like your brand (not a generic bot).
How This Automation Works
The full n8n workflow, from trigger to final output:
n8n Workflow Template: WhatsApp + OpenAI: brochure based replies that sell
flowchart LR
subgraph sg0["WhatsApp Flow"]
direction LR
n0["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/whatsapp.svg' width='40' height='40' /></div><br/>WhatsApp Trigger"]
n1@{ icon: "mdi:brain", form: "rounded", label: "OpenAI Chat Model", pos: "b", h: 48 }
n2@{ icon: "mdi:memory", form: "rounded", label: "Window Buffer Memory", pos: "b", h: 48 }
n3@{ icon: "mdi:wrench", form: "rounded", label: "Vector Store Tool", pos: "b", h: 48 }
n4@{ icon: "mdi:vector-polygon", form: "rounded", label: "Embeddings OpenAI", pos: "b", h: 48 }
n5@{ icon: "mdi:brain", form: "rounded", label: "OpenAI Chat Model1", pos: "b", h: 48 }
n12["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/whatsapp.svg' width='40' height='40' /></div><br/>Reply To User"]
n13["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/whatsapp.svg' width='40' height='40' /></div><br/>Reply To User1"]
n14@{ icon: "mdi:memory", form: "rounded", label: "Product Catalogue", pos: "b", h: 48 }
n16@{ icon: "mdi:swap-horizontal", form: "rounded", label: "Handle Message Types", pos: "b", h: 48 }
n17@{ icon: "mdi:robot", form: "rounded", label: "AI Sales Agent", pos: "b", h: 48 }
n17 --> n12
n0 --> n16
n4 -.-> n14
n1 -.-> n17
n14 -.-> n3
n3 -.-> n17
n5 -.-> n3
n16 --> n17
n16 --> n13
n2 -.-> n17
end
subgraph sg1["When clicking ‘Test workflow’ Flow"]
direction LR
n6@{ icon: "mdi:play-circle", form: "rounded", label: "When clicking ‘Test workflow’", pos: "b", h: 48 }
n7@{ icon: "mdi:vector-polygon", form: "rounded", label: "Embeddings OpenAI1", pos: "b", h: 48 }
n8@{ icon: "mdi:robot", form: "rounded", label: "Default Data Loader", pos: "b", h: 48 }
n9@{ icon: "mdi:robot", form: "rounded", label: "Recursive Character Text Spl..", pos: "b", h: 48 }
n10@{ icon: "mdi:cog", form: "rounded", label: "Extract from File", pos: "b", h: 48 }
n11["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>get Product Brochure"]
n15@{ icon: "mdi:memory", form: "rounded", label: "Create Product Catalogue", pos: "b", h: 48 }
n10 --> n15
n7 -.-> n15
n8 -.-> n15
n11 --> n10
n9 -.-> n8
n6 --> n11
end
%% Styling
classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef disabled stroke-dasharray: 5 5,opacity: 0.5
class n0,n6 trigger
class n17,n8,n9 ai
class n1,n5 aiModel
class n3 ai
class n2,n14,n15 ai
class n4,n7 ai
class n16 decision
class n11 api
classDef customIcon fill:none,stroke:none
class n0,n12,n13,n11 customIcon
The Problem: WhatsApp Questions Turn Into Manual Brochure Hunting
WhatsApp is where buyers actually ask. It’s also where accuracy quietly falls apart. One person answers from the latest brochure, another answers from memory, and a third forwards a screenshot that’s already outdated. Meanwhile, every “Does it come in size X?” or “Is it compatible with Y?” forces you to reopen the PDF, search, skim, copy, then rewrite so it sounds human. Multiply that by 20 chats a day and you’ve burned a chunk of the workday on tasks that don’t move revenue forward. And honestly, the worst part is the mental load. You’re context-switching all day.
It adds up fast. Here’s where it usually breaks down in real teams.
- Replies get slower as soon as more than one person is handling the inbox.
- Details drift over time, which means customers receive different answers to the same question.
- Copy-pasting from PDFs leads to clunky messages, so you end up rewriting everything anyway.
- Non-text messages (voice notes, images) derail the flow and create messy handoffs.
The Solution: WhatsApp Answers Pulled Directly From Your PDF
This workflow turns your WhatsApp number into a lightweight AI sales agent that answers questions using your actual brochure content, not guesswork. You start by pointing n8n at a publicly accessible PDF (your product brochure). The workflow downloads it, extracts the text, splits it into readable chunks, and creates OpenAI embeddings so the content becomes searchable. Those embeddings are stored in an in-memory vector store so the AI agent can retrieve the right section when a customer asks something specific. When a WhatsApp message arrives, the workflow checks the message type. If it’s text, the AI agent generates a natural reply grounded in the brochure. If it’s not text, the workflow politely responds that the message type isn’t supported (so your inbox stays clean).
The flow starts with a WhatsApp webhook trigger and a quick filter for message types. Then OpenAI retrieves brochure context from the vector store and drafts the response. Finally, the workflow sends the finished answer back to the same WhatsApp chat, ready to keep the conversation moving.
What You Get: Automation vs. Results
| What This Workflow Automates | Results You’ll Get |
|---|---|
|
|
Example: What This Looks Like
Say your team gets 30 WhatsApp product questions a day. Manually, you might spend about 3 minutes finding the brochure section, then 2 minutes rewriting it into a friendly message, so roughly 5 minutes per chat (about 2.5 hours daily). With this workflow, the “work” is basically zero after setup: the message triggers instantly, the AI drafts a brochure-based reply in about a minute, and it sends automatically. You still review tricky edge cases, but you’re no longer doing the brochure scavenger hunt.
What You’ll Need
- n8n instance (try n8n Cloud free)
- Self-hosting option if you prefer (Hostinger works well)
- WhatsApp Business API connection for receiving and sending messages
- OpenAI to generate embeddings and replies
- OpenAI API key (get it from the OpenAI dashboard)
Skill level: Intermediate. You’ll be pasting credentials, updating a PDF URL, and running a one-time “build the knowledge base” test execution.
Don’t want to set this up yourself? Talk to an automation expert (free 15-minute consultation).
How It Works
A WhatsApp message triggers the workflow. The WhatsApp trigger receives the incoming chat and passes the message details into n8n through a webhook.
Message type is checked. A Switch/If block filters for text messages only. If someone sends audio, an image, or a document, the workflow routes to a simple “unsupported message type” reply so your bot doesn’t hallucinate.
Your brochure becomes searchable. On the one-time setup run, n8n downloads the brochure PDF (HTTP Request), extracts the text, chunks it, and generates embeddings using OpenAI. Those embeddings are stored in the in-memory vector store so the AI agent can retrieve relevant passages later.
The AI agent drafts the response and sends it back. The AI Agent uses the OpenAI Chat Model plus the retrieved brochure context to write a factual reply, then a WhatsApp “send message” node posts it back to the same conversation.
You can easily modify the brochure source to use Google Drive files instead of a public URL based on your needs. See the full implementation guide below for customization options.
Common Gotchas
- WhatsApp Business API credentials can expire or need specific permissions. If things break, check your WhatsApp provider dashboard and the n8n credential test first.
- If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
- Default prompts in AI nodes are generic. Add your brand voice early or you’ll be editing outputs forever.
Frequently Asked Questions
About 30–60 minutes if your WhatsApp and OpenAI accounts are ready.
No. You’ll mostly paste credentials, update your PDF URL, and tweak the AI prompt.
Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI API costs, which are usually a few cents per conversation depending on brochure size and traffic.
Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.
Yes, but you’ll want to swap the single PDF download step for a file picker or routing logic. Many teams store several brochures in Google Drive and choose which one to embed based on a keyword, a menu reply, or the customer’s first message. You can also move from the in-memory vector store to a persistent store like Qdrant or Pinecone so multiple catalogs are always available. The simplest customization is editing the system message inside the AI Sales Agent to define which product lines it should focus on and what it should refuse to answer.
Usually it’s expired credentials or a webhook mismatch. Confirm the WhatsApp Trigger webhook URL matches your n8n domain, then re-save credentials in both the WhatsApp trigger and WhatsApp send nodes. If you’re using a provider that requires approved templates, make sure your outbound message format is allowed. Rate limits can also show up as random failures when volume spikes.
On a typical n8n Cloud plan, it can comfortably handle hundreds of chats a day for most small teams, and self-hosting scales based on your server.
Often, yes, because this is more than a simple trigger-and-send Zap. You’re doing document processing, retrieval from a vector store, and conditional routing, which tends to get expensive or awkward in tools built for linear automations. n8n also gives you the self-host option, which matters once message volume grows. Zapier or Make can still be fine for basic auto-replies, but brochure-grounded answers are a different category. If you want help deciding, Talk to an automation expert.
Set this up once, and your WhatsApp inbox stops depending on who’s online and who remembers the brochure details. The workflow handles the repetitive questions so your team can focus on real sales conversations.
Need Help Setting This Up?
Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.
