WhatsApp MCP Use Cases: 9 Ways Teams Run WhatsApp from Claude or ChatGPT
Nine things your team can do once its WhatsApp inbox is connected to Claude, ChatGPT or Cursor, each with the prompt to copy and a short lesson that shows it working.
9 min read
Your team's WhatsApp holds the answers to most questions a manager asks on a Monday morning. Who is waiting on a reply? Which deal went quiet? What did we promise this customer last week? Finding out usually means scrolling through every chat on every number.
WhatsApp MCP by TimelinesAI connects your team's WhatsApp inbox to the AI assistant you already use: Claude, ChatGPT, Cursor, Codex or Grok. You ask in plain language and the assistant reads the conversations. When you tell it to, it also replies, labels, assigns and closes chats for you.
This page covers nine things teams use it for, with the prompts to copy. Each one links to a short lesson in the MCP Academy that shows it working on a demo workspace.
Key takeaways
- WhatsApp MCP is about visibility and control. Ask what's waiting, who's behind and what was said, across every number your team uses.
- Nothing changes until you ask. Most of these use cases only read your inbox. Anything that changes a chat or reaches a customer starts as a plan or a draft you approve.
- Changes show up in TimelinesAI straight away, exactly as if a teammate had made them. Your team keeps working there as before.
- It works with regular WhatsApp numbers and the WhatsApp Business API. Numbers connected by QR code need no Meta approval.
- Included in every TimelinesAI plan at no extra charge. Messages sent through it use your normal messaging quota.
What WhatsApp MCP is, in one minute
MCP (Model Context Protocol) is the open standard AI assistants use to connect to other apps. TimelinesAI's MCP server connects your assistant to your team's real WhatsApp inbox in TimelinesAI: every number, every teammate, every chat.
Once it's connected, your assistant gets 34 tools. With them it can:
- Read: list and filter chats (unread, open, by teammate, by label, groups or direct chats), read a conversation's history, and check your messaging quota.
- Act: reply, react, label, assign, open and close chats.
- Work with the WhatsApp Business API: see which chats are still inside the service window, list your approved templates and send them.
And a few things it isn't:
- It isn't a chatbot. It doesn't answer your customers on its own. It works for you and your team, inside the assistant you already use.
- It isn't a new inbox. Labels, owners and replies the assistant sets appear in TimelinesAI, where your team already works.
- It doesn't get extra permissions. Every action runs as you, with your role's permissions in TimelinesAI.
The nine use cases at a glance
| Use case | Best for | What you ask |
|---|---|---|
| 1. Morning briefing | Managers | "Give me a morning briefing: unread by teammate, anything waiting more than two business days…" |
| 2. Who's waiting, and on whom | Team leads | "Which open chats have a customer message older than two business days with no reply?" |
| 3. Cover for a colleague | Anyone taking over chats | "Summarize every chat with this contact from this month." |
| 4. Prep for a call | Sales reps | "What is the customer asking for, and what did we promise?" |
| 5. Follow up quiet HubSpot deals | Sales teams on HubSpot | "Find my open deals with no activity in the last 7 days…" |
| 6. Spot buying signals | Sales leads | "Who asked about price or a discount this week?" |
| 7. Triage the shared inbox | Support teams | "Propose labels and an owner for each unread chat." |
| 8. Reply faster, draft first | Support and sales | "Draft a reply to this chat. Don't send yet." |
| 9. Business API window and templates | Teams on the WhatsApp Business API | "Which chats still have an open service window?" |
For managers and team leads
A morning briefing before your first call
Many managers spend the start of the day working out what happened overnight. Ask for it instead:
Give me a morning briefing: unread by teammate, anything waiting more than two business days, and unassigned chats from overnight. Keep it to one screen.
Then narrow it to what you own:
Same briefing, but only for the sales number.
Save the prompt and reuse it every morning. Telling the assistant the shape you want, such as "one screen" or "most urgent first", matters as much as telling it what to look for.
Learn it: What needs me today › Build your briefing
See which customers are waiting, and on whom
Slow replies are one of the most common reasons WhatsApp leads go cold. These prompts only read your inbox, so they're safe to run at any time:
List my unread open WhatsApp chats, grouped by the teammate responsible for each. Don't open or change anything.
Which open chats have a customer message older than two business days with no reply from us? Oldest first.
Show chats that arrived since 7pm yesterday and have no one assigned.
Groups work the same way. Add "groups only" to any of these prompts to check the WhatsApp groups you share with clients.
Learn it: What needs me today › Find who's waiting
Cover for a colleague without losing the thread
When a chat changes hands, from sales to support or to someone covering a holiday, the new owner needs the story fast. The assistant reads the whole thread, including messages sent from the WhatsApp app on a phone:
Summarize every chat with [contact] from this month, oldest to newest.
Then hand the chat over:
Assign chat [chat ID] to [teammate name].
The assistant looks your teammate up first, so the chat lands with the right person, and the customer never has to repeat themselves.
Learn it: Understand a conversation › Summarize a thread
For sales teams
Prep for a call in a minute
Deals turn on details: what the customer asked for, at what price, and what your side promised. Name the fields you need and they come back in the same order every time:
From chat [chat ID], tell me: what the customer is asking for, any price or quantity mentioned, and what was promised by our side.
Based on chat [chat ID], what's the next step, and who owns it?
Treat the answer as a map. Before you quote a price or a date back to the customer, ask to see the original message:
Show me the conversation around the last message where the customer mentioned a delivery date in chat [chat ID].
Learn it: Understand a conversation › Pull out what matters
Follow up the HubSpot deals that went quiet
HubSpot knows which deals stalled. TimelinesAI knows what was last said on WhatsApp. Add HubSpot's own connector to the same assistant and one question can reach both:
Find my open HubSpot deals with no activity in the last 7 days. Show the deal name, stage, contact and phone number.
For each of those contacts, find the matching WhatsApp chat in TimelinesAI and tell me when we last messaged them.
Draft a short, friendly WhatsApp follow-up for each deal, based on its stage and our last WhatsApp conversation. Number the drafts. Don't send anything.
Read the drafts, then send only the ones you approve, for example "Send drafts 1 and 3." The assistant matches HubSpot contacts to WhatsApp chats by phone number, so keep numbers in the same international format in both places. TimelinesAI's HubSpot sync keeps running exactly as before.
Learn it: HubSpot and WhatsApp together. Full setup: HubSpot WhatsApp MCP guide
Spot buying signals across this week's chats
Sales leads want to know who is ready to buy without reading every conversation:
Go through this week's open chats on the sales number. List every customer who asked about price, a discount or a delivery date, one line each, with the chat ID.
The assistant reads the chats one by one, so a tight scope gives a faster and more complete answer: one number, one week, open chats only. On a busy inbox, run it once per teammate.
Want something you can install and reuse? The WhatsApp MCP skills include ready-made recipes for briefs before a call and for chasing chats that went quiet.
For support teams and shared inboxes
Triage the inbox: plan first, then act
Labels and owners stop a shared inbox from becoming nobody's job. Let the assistant propose them, then approve only what you agree with:
Look at my unread chats. Propose labels and an owner for each, as a numbered list. Don't change anything yet.
Apply 1, 2 and 4.
When a chat is handled, close it. Closing moves it out of the active inbox without deleting anything, and you can reopen it at any time:
Close chat [chat ID]. It's handled.
Learn it: Label, assign, reply › Plan, then act
Reply faster, with a draft you approve
Messages the assistant sends come from your WhatsApp number, and the customer can't tell them apart from ones you typed. So start every reply as a draft:
Draft a reply to chat [chat ID] based on the conversation. Show me the draft. Don't send yet.
Change anything you like, then say "Send it." For a quick acknowledgement, a reaction is often enough:
React 👍 to the last message in chat [chat ID].
Replies sent this way use the same monthly messaging quota as the TimelinesAI app. Before replying to more than a few chats, ask the assistant how much messaging quota you have left.
Learn it: Label, assign, reply › Reply, draft first
For WhatsApp Business API numbers
Work inside the service window, and use templates outside it
On the WhatsApp Business API, you can send free text only within 24 hours of the customer's last message. After that, and to start a new conversation, you need a template that Meta has approved. The assistant can tell which chats are still open:
Which of my WhatsApp Business API chats still have an open service window? Draft replies for those first.
List our approved WhatsApp templates.
The service window on chat [chat ID] has closed. Send our [template name] template in English. Show me first.
Why did my last template message to this customer fail?
Meta charges for each template message, so confirm before the assistant sends one. Not sure which kind of connection suits your team? Compare QR, Coexistence and the WhatsApp Business API.
Three rules that keep it safe
- Read first. Start with the questions that only read. Adding "don't change anything" to a prompt keeps it that way while you learn.
- Plan or draft, then approve. For anything that changes a chat or reaches a customer, ask for a numbered plan or a draft, check it, then say what to apply. Sends are real and immediate, and there's no undo.
- Keep sends small. Follow up a handful of chats at a time. WhatsApp can ban numbers that send cold outreach or broadcasts, so don't use the assistant for mass messaging.
Which AI assistants work with it
- Claude: web, desktop and Claude Code
- ChatGPT: through the TimelinesAI plugin
- Cursor, Codex and Grok Build
- Any other assistant or agent that supports remote MCP servers
You sign in once with your TimelinesAI account. There are no keys to copy.
Get started in three steps
Step 1: Connect your WhatsApp number to TimelinesAI
Start a 10-day free trial, with no credit card needed, and scan the QR code with the WhatsApp number your team uses. WhatsApp Business API numbers work too.
Step 2: Add TimelinesAI to your assistant
- Claude: open Customize → Connectors, click +, then Add custom connector. Name it TimelinesAI, paste
https://mcp.services.timelines.ai/mcpand click Add. On a Team or Enterprise plan, an Owner adds it once under Organization settings → Connectors. - ChatGPT: open the TimelinesAI plugin and click Install plugin.
- Cursor, Codex, Grok Build and other clients: follow the setup tabs on the WhatsApp MCP page.
- Claude: open Customize → Connectors, click +, then Add custom connector. Name it TimelinesAI, paste
Step 3: Sign in and ask your first question
A browser tab opens. Sign in with your TimelinesAI account and approve the connection. Then start with a question that only reads:
PromptList my unread open WhatsApp chats, grouped by the teammate responsible for each. Don't open or change anything.
FAQ
What is WhatsApp MCP?
WhatsApp MCP by TimelinesAI connects your team's WhatsApp inbox to an AI assistant such as Claude, ChatGPT or Cursor, using the open Model Context Protocol. The assistant can read your chats and, when you ask, reply to, label, assign and close them.
Does WhatsApp MCP cost extra?
No. WhatsApp MCP is included in every TimelinesAI plan at no extra charge. Messages sent through it use the same monthly messaging quota as the TimelinesAI app. See pricing for the plans.
Is WhatsApp MCP a chatbot?
No. A chatbot answers your customers automatically. WhatsApp MCP gives you and your team an assistant that works on your inbox: it finds what's waiting, summarizes conversations and drafts replies, and it sends only when you tell it to.
Can the assistant send messages to my customers?
Yes, when you ask it to. Messages go out from your WhatsApp number straight away, and the customer can't tell them apart from ones you typed. That's why every use case on this page asks for a draft first.
Do I need the WhatsApp Business API?
No. TimelinesAI connects your existing WhatsApp number by QR code, with no Meta approval process. If you already use the WhatsApp Business API, those numbers work too.
Does it work with WhatsApp groups?
Yes. The assistant can list group chats separately from direct chats, so you can ask about the groups you share with clients in the same way.
Which AI assistants can I use?
Claude (web, desktop and Claude Code), ChatGPT, Cursor, Codex, Grok Build and any other client that supports remote MCP servers.
Does it work with my CRM?
TimelinesAI keeps syncing WhatsApp to your CRM as it does today. To ask one question across both, add your CRM's own connector to the same assistant. We tested this with HubSpot and wrote up the setup and prompts.
Where to go next
- MCP Academy: five short lessons, free to read, no signup
- WhatsApp MCP skills: ready-made recipes for triage, call briefs and follow-ups
- WhatsApp MCP docs: every tool, with example prompts
- Analyze WhatsApp with Claude Code: response times and topic trends across weeks of chats (this kit uses the TimelinesAI public API)
- WhatsApp MCP: what's included and how to connect each assistant