A chatbot should hand a conversation over to a person when the customer explicitly asks to talk to someone, expresses a complaint or anger, brings up returns or money, isn't understood on the same question twice, or needs a personalized decision. In a well-designed live chat chatbot setup, the bot answers routine questions instantly and, at critical moments, passes the conversation to your team along with its full history. That way, customers get fast answers and also get to deal with a real person on the issues that matter.
Before you set up this balance, it helps to take a look at our basic guide on the difference between chatbots and live chat.
What is live chat, and which part does a chatbot take over?
Live chat is when a customer messages in real time with a business's employee over messaging channels. The live chat team is the people who handle these conversations. At small businesses, this team is often the owner or one or two employees.
The problem is that most incoming messages revolve around the same few topics: price, address, business hours, stock and delivery time. Having an employee answer these questions one by one eats up the time that should go to conversations that genuinely need a human touch. That's exactly where a chatbot comes in:
- It answers repetitive information questions 24/7. It handles many conversations at the same time. It collects information from the customer, such as their name, phone number and request. It carries the conversation up to the point where a person needs to take over.
In other words, the goal isn't to take people out of the loop, but to direct them to the right conversations.
When should a chatbot hand off to a human? 7 clear signals
1. When the customer explicitly asks for a person
Stalling a customer who writes "I want to talk to a customer service representative" with a bot damages trust. When this request comes in, the bot shouldn't insist; it should transfer the conversation and tell the customer when they'll hear back.
2. When there are signs of a complaint or anger
With complaints like a damaged product, a delayed order or the wrong service, customers want to be understood. The bot can apologize and collect information in the first message, but an employee should take on the resolution.
3. When returns, money or exceptions come up
Approving a return, issuing a refund, giving a special discount or granting an out-of-policy request all require making a decision on behalf of the business. The bot shouldn't make these decisions on its own. We explain in detail how these boundaries are drawn in e-commerce in our post on the limits of an e-commerce chatbot.
4. When the bot fails to understand the same question twice
If the customer keeps rephrasing their question and the bot can't give the right answer, it's better to transfer the conversation than to keep the loop going. These chats also reveal topics that are missing from the bot's knowledge source.
5. When a high-value sales opportunity appears
A customer negotiating a bulk order, a corporate quote or an expensive product benefits from talking to a sales representative. The bot can collect preliminary information and pass a warm lead to the team.
6. When a personal assessment is required
Topics like health, legal matters, finance or technical faults require personalized advice. The bot can give general information, but an expert should make the assessment.
7. When sensitive personal information is shared
If a customer starts sharing their personal situation or sensitive information in the chat, it's better for an employee to handle the conversation carefully.
Which topics should the bot handle, and which should a person?
Topic | Bot | Person |
|---|---|---|
Price, product and service information | Answers | Only for negotiations or special offers |
Address, business hours, directions | Answers | Not needed |
Appointment and reservation requests | Collects the details | Confirms |
Order status | Explains the general process, takes the order number | Steps in if there's a problem |
Complaints, damage, returns | Gives the first reply, collects information | Resolves and decides |
Corporate or bulk requests | Collects preliminary information | Prepares the quote |
What does a good handoff look like?
Handing off to a person isn't just saying "I'm transferring you." If the handoff is done poorly, the customer has to explain everything from the start, and the time the bot saved is lost. Follow these steps for a smooth handoff:
- Let the customer know: The bot should state clearly that the conversation is being passed to the team.
- Pass along the history: The employee should be able to see what the customer wrote to the bot so they don't ask the same questions again.
- Alert the team instantly: When a handoff happens, the person responsible should get a notification. If you're away from your desk, you can join the conversation quickly by following chats from the mobile app.
- Set expectations: If the handoff happens after hours, the bot should collect the customer's details and state when they'll hear back.
- Log the outcome: Noting the result when the conversation ends helps you improve the bot for similar situations.
Common mistakes in handoff rules
The problems that upset the balance are usually not technical but come from how things are designed. These are the most common mistakes:
- Forcing the bot to answer every question: A bot that guesses on topics that aren't in its knowledge source starts giving wrong information. It's safer for it to transfer the conversation on topics it doesn't know.
- Handing off too early: A bot that leaves every conversation to a person by the second message doesn't reduce the team's workload. Information questions should stay with the bot.
- Responding late to handed-off conversations: If a customer waits for hours after the "I'm transferring you" message, it feels as if the bot never existed. Set an internal response target for the conversations you hand off.
- Leaving ownership unclear: If the live chat line depends on a single person, conversations are left without an owner when that person is on leave. Assign a backup.
How to balance live chat and a chatbot with Chatbu
Chatbu's approach can be summed up as "automation first, people at the moments that matter." You can see how you set the bot's tone, routing logic and rules for stepping in on the page that explains the balance between automation and people in Chatbu. Here are the setup steps:
- Train the bot with your own data: Add your web pages and files like your price list and return terms. Leave outdated content out of training.
- Define the behavior: Specify which topics the bot should answer and in which situations it should leave the conversation to the team. Customize its brand voice, behavior and design to fit your business.
- Use live handoff: According to Chatbu's features page, when a visitor wants to speak to a real person, the chat is passed to your team along with its full history, and your team is notified.
- Connect your channels: Embed the bot on your website, and connect your Instagram DM and WhatsApp accounts through the Meta integration.
- Turn on lead collection: In conversations with sales potential, collect the name and phone number in the chat.
- Install the mobile app: Follow live chats from your phone, get notifications and join conversations when needed.
Measure the balance with chat history
During initial setup, you decide which topics go to a person based on educated guesses. A few weeks later, the chat history shows you the reality. Look at these questions:
- Which topics led to the most handoffs? Which handed-off conversations were actually information questions the bot could have answered? On which questions did the bot fail to satisfy the customer?
Add topics the bot could answer but handed off because of missing information to the knowledge source. For topics that genuinely require a person, make the handoff rules clearer. This cycle ensures your live chat team only deals with the conversations where they add value.
Frequently Asked Questions
What is live chat?
Live chat is when a customer messages in real time with a business's employee over channels like the website, Instagram or WhatsApp. A chatbot takes over the routine part of this process and lightens the team's workload.
What should I say to get connected to a customer service representative?
On the business side, the answer to this question should be simple: when the customer writes "I'd like to talk to a representative" or something similar, the bot should understand the request and transfer the conversation. Don't force customers to memorize a special command.
How do you provide live chat support?
Bring your messaging channels together, assign people responsible, leave frequently asked questions to the bot and respond quickly to handed-off conversations. Mobile notifications make it easier to reply even when you're away from your desk.
Will a chatbot replace the live chat team?
No. A chatbot answers repetitive questions and collects information. Complaints, returns, negotiations and personalized decisions are still your team's job.
How does handoff work after hours?
After hours, the bot collects the customer's request and contact details and states when they'll hear back. When business hours start, your team sees the conversation with its history and picks up where it left off.
Leave routine questions to the bot and the moments that matter to your team. Try Chatbu free for 14 days, no payment details required. To get started, visit chatbu.io.
