Workflows
WhatsApp customer support, from first message to resolved
Good support on WhatsApp is less about clever bots and more about never losing a chat. This page walks one customer question from first message to resolution and shows which part each Samvadify tool plays.
Create account
How it works
From first message to resolved
- 01
A keyword rule or the AI agent answers the common question straight away
- 02
If the customer needs a person, the chat is handed over and assigned
- 03
The agent replies, sets a status and adds notes to the contact
- 04
The agent marks the chat resolved and the next waiting customer is assigned
Step 1: answer the obvious questions instantly
Opening hours, delivery areas, price lists and “how do I book” make up a large share of most support inboxes. Keyword auto-reply rules give a fixed answer the moment those words appear, with no AI credit used.
For questions that are worded in many different ways, you can add an AI agent trained on your own instructions. AI replies are billed to the owner’s AI wallet, so keep fixed rules for anything that never changes.
Step 2: hand off to a person cleanly
The AI agent can transfer a chat to a human. Handovers start with the most junior agent level that is online, and that agent can escalate upward from the inbox. While a person has the chat, the AI stays out of it.
Agents set themselves online or offline. When they come online, waiting customers can be assigned to them automatically, and owners can also assign or unassign chats by hand.
- Tell the customer a person is taking over.
- Do not let a bot keep answering after the handoff.
- Keep the handoff rule simple: refunds, complaints and anything the bot is unsure about.
Step 3: work the chat in a shared inbox
Every conversation has a status: open, needs attention, follow up, pending, waiting for customer, resolved or spam. Sorting by status keeps urgent chats at the top and stops two people replying to the same customer.
The CRM profile sits beside the chat, so notes, tags and deal stage from earlier conversations are visible. Tasks can be created for anything that needs doing later.
Step 4: close the loop
Resolving a chat hands the thread back to automation, frees the agent and assigns the next waiting customer. A handovers report shows how many chats moved to people and can be exported as a CSV.
Remember the 24-hour customer service window. If you need to come back to a customer after it closes, you need an approved template and their permission.
What is not included
Samvadify does not publish guaranteed response times, and there is no built-in customer satisfaction survey. Ticketing tools outside WhatsApp are not integrated natively; use webhooks if you need to pass chats to another system.
Setting it up in a day
Start with your five most common questions and write a fixed reply for each. Invite your team, decide who handles which hours, and agree a short handoff rule. Add the AI agent only after the fixed replies are working, and review the handovers report at the end of the first week to see what customers still ask people for.
Illustrative example
A delayed order
Customer: My order hasn’t arrived, it’s been 6 days.
AI agent: Sorry about that. I’m passing you to our team so they can check your order.
Agent: Hi, I’m Ritu. Could you share your order number? I’ll check with the courier now.
Agent (later): It’s out for delivery today. I’ve marked this resolved, but reply here if it doesn’t arrive.
Common questions
Do I need the AI agent for support?
No. Keyword auto-replies and a shared inbox work on their own. The AI agent is optional and billed to your AI wallet.
Can several agents answer from one number?
Yes. Team members log in separately, chats are assigned to them, and seat limits depend on your plan.
What happens when a chat is resolved?
The thread goes back to automation, the agent is freed and the next waiting customer can be assigned.
Can I message a customer days later?
Only with an approved template and the customer’s permission, because the 24-hour service window will have closed.
Related reading
See also: an AI agent that hands chats to people; automation rules for repeat questions.