Chatbot Support: When Should a Human Take Over?
Chatbot support greatly facilitates the routine of customer service in any company. But it by no means replaces human support.
And there’s no point in trying. When it comes to customer service, it’s always better to start with a chatbot and, if necessary, proactively offer human support.
This is because the chatbot organizes the queue of service requests while, if possible, already trying to resolve users’ doubts.
The topic of today’s article is this: up to what point should chatbot support go? And how can we make the transition to human support in the best possible way?
Let’s better understand this together. And let’s also understand the other main type of chatbot available on the market: the one designed to generate leads.
Starting now:
Understanding the Two Main Types of Chatbots Available on the Market Today
Chatbots are very versatile tools and can be used for various purposes.
Today, in the market, the two main purposes for chatbots are customer service and lead generation.
Customer service is usually done through specific platforms, which can offer direct support on the website, via SMS, WhatsApp, and Telegram.
Meanwhile, the focus of lead generation chatbots is a bit broader: they work directly on your website to automatically qualify leads.
It’s interesting to better understand how these two types of chatbots work to be clear about what to expect from them—how far they can go and what they usually offer.
Let’s understand better together!
Chatbot for Customer Service
Customer service chatbots work in a way that’s easy to understand.
They can be configured in various ways and have different levels of functionality depending on the contracted platform.
Rest assured that, throughout this article, we’ll discuss much more about chatbot customer service and all the differences that platforms usually have.
But to summarize, chatbot support aims to offer simple options for resolving basic problems for the user.
And a bit beyond that too: it also seeks to provide a conversational approach from the first contact, even if the request hasn’t been resolved and it’s necessary to talk to a human.
Many requests are simple to resolve and can be done directly through the chatbot. Banks, for example, offer most of their functionalities through customer service via chatbots.
Functionalities, in this case, for problem resolution and requests that you previously could only resolve at the cashier’s desk or by talking to your manager.
This is the main role of chatbot customer service. But what is its counterpart? How does the lead generation chatbot work?
Let’s discuss them better in the topic below. Come with me:
Lead Generation Chatbots
The lead generation chatbot works a bit differently. Its main concern is converting visitors into leads for your website.
These chatbots are the natural evolution of forms on Landing Pages.
In the case of these chatbots, there is no need for a human to provide any type of support. In fact, there is no support for human assistance.
All the work of the lead generation chatbot is automated; this is its main differential compared to other types of chatbots.
What Are the Types of Customer Service Chatbots?
There are very different customer service chatbots, mainly in the features they offer.
The overall design of this type of tool doesn’t change much. They all operate quite similarly in customer service.
But it’s their functionalities, what’s under the hood, that makes all the difference.
This difference, in fact, impacts the very hiring of the tool.
Let’s now understand the types of chatbots according to the features they offer.
If you want to delve deeper into the topic, I suggest our more comprehensive article, which covers 10 types of chatbots in detail. Access it below:
➡️ 10 Types of Chatbot and How To Choose the Best
So let’s learn more about the types of customer service chatbots?
Rule-Based Chatbots
These are the simplest types of chatbots to configure.
They are chatbots that do not independently converse with the user. Their operation is based on simple flows and pre-determined rules.
This is the case with WhatsApp chatbots, for example.
Being a simpler chatbot, the rule here is clear: chatbot support should go as far as the chatbot can handle.
There are queries that are too complex for this type of chatbot. Most of the time, it operates more simplistically, resolving basic requests with the support of the company’s systems.
This is the case with bank chatbots, which we mentioned above. WhatsApp requests use this model to resolve practical issues, such as issuing duplicate invoices, increasing limits, etc.
Normally, this chatbot operates in two distinct ways:
- Choice of options: This is when the user only chooses the answers they can give within the flow. Leadster’s chatbot works this way, for example. In our case, it works because there is no need to have an extensive conversation with the chatbot, which operates as a CTA throughout the website;
- Keyword chatbot: Here, the user can maintain a certain conversation with the chatbot, which doesn’t fully understand it but can identify certain keywords and deliver related options to the user. You say something like “I missed the payment date, I need a duplicate,” and the chatbot understands the keyword “duplicate” and responds accordingly;
Contextual Chatbots
These are more advanced chatbots that understand the context of the message and offer options according to what the user is asking.
These are often understood as true virtual assistants. The closest and most advanced examples we have include Alexa, Siri, Cortana, etc.
Contextual chatbots also have levels of complexity. Some can conduct the conversation naturally, while others only offer options based on what is being said.
This is the true future of chatbots: conversing with them is like conversing with a human.
However, there is still no contextual chatbot that can think and act like a human in customer service situations.
For this to happen, technology is needed that can understand what the user is saying and deliver correct responses based on the company’s database without losing compliance.
We are well advanced in this first need, but in the second, we are still developing tools.
The closest we can get today are chatbots from Zendesk and Salesforce, which capture information from requesters, understand what they need, and offer resolution suggestions to the human attendant.
Until When Should Chatbot Support Go?
Well, so far we’ve talked about the types of chatbots and everything they can deliver within a customer service strategy.
Although there are not many customer service chatbots that do all the work, in many cases, it is not necessary to involve human attendants.
There are requests that are so simple that human support will actually hinder rather than help.
But there are also cases where chatbot support needs to be as quick as possible and transferred to humans with agility.
Let’s discuss these cases now. When is the ideal moment to pass the baton?
And right after, let’s talk about the best ways to make this transfer.
Come with me:
Moments When the Chatbot Can Operate Alone
First, it’s worth discussing when the chatbot can operate alone and there’s no need to pass the support to a human for now.
Normally, it’s important to keep the user’s expectation as the main focus. There are times when human support hinders rather than helps.
I made a quick scheme here to better understand which moments these are:
- Simple requests: When it is possible to resolve the request through the chatbot without involving a human, it’s better to resolve it this way. Avoid complicating what is simple. Duplicate invoices, delivery tracking, package tracking: all of this can be resolved by the chatbot alone with integrations into your systems;
- Also simple information: Freight calculation, pricing of products and services, tracking the status of a request—all of this can also be resolved by the chatbot;
- Proactive messages: When your system goes down, service instability issues, etc.: it’s important to use the customer service chatbot so that, right from the first conversation, it informs about the problem and the resolution;
- Direct interactions with the system: When your client needs to send documentation, for example, which will be stored directly in the system. Or when they need that documentation. The chatbot itself can deliver what is needed, depending on the technology. But it’s important to understand here that, in the case of a chatbot that does not offer this type of support, a human attendant needs to be activated.
Moments When It Is Important for a Human to Provide Support
So we understand that on some occasions, it’s even better for the support work to be conducted solely by the chatbot, right?
As we saw, most of these situations involve simple requests, where the process of human support will be more time-consuming than resolution by the chatbot.
Think for a moment about the path your user needs to take to access the chatbot: they need to get in touch, then select their options, then receive the chatbot’s response.
And in this process, the user decides if they need more requests to be resolved.
When you involve a human in the process, more steps arise: they need to ask to speak with a human agent, then wait for the agent, and only then does the support begin again.
And depending on your company’s chatbot system, it’s necessary to consider that the chatbot might not have the option to save the requester’s information, making the agent ask for everything again.
But this is rare. The main point to consider is what is easier for the requester at the moment: speaking with the chatbot or with the human?
Below are some situations where it is important for the support to be transferred. Learn more:
When Asking Sensitive Questions
If you are going to ask more in-depth questions about the requester, such as sensitive personal data, it is better that the option to request a human agent is available.
Or that you automatically make the transition, informing the user that it will happen.
Many people are suspicious when delivering very detailed personal documents, copies of these documents, etc.
So, it’s better for the user to know that if needed, an agent can explain everything at any moment.
When Dealing with Files
Chatbots can have complications when working with files that the user needs to send.
Copies of contracts, personal documents, etc. Even if the documents are not very sensitive, it is better for their receipt and upload into the system to be done by human agents to avoid errors.
In Negotiations
When negotiating debts or overdue fees, it is also better for the support to be provided by humans.
This is because you have a much better chance of resolving the delinquency when the collection is made by a person and not a robot.
Human agents can negotiate conditions that go beyond the rules determined by the chatbot.
And these negotiations can be simple. For example: you are dealing with a delinquent client who has six overdue installments, something common in private schools.
During the resolution of the case, the client asks the school to generate an invoice with a due date of seven days.
The problem: the rule established in the chatbot only creates invoices due on the day of the request. If the support is only done by the chatbot, you might frustrate the delinquent client and lose the negotiation.
A human agent can ask for authorization from the financial manager and issue the invoice as an exception.
In Complex Requests
This is the most common case in chatbot support: users leave the chatbot to make the primary qualification of the client’s request. What they need, their identification, etc.
Then, the chatbot presents options to the user.
If the request is very complex, it is important to guide the user to a human agent immediately.
The work is easier with chatbots that understand keywords. Thus, it can determine if the support is simple or complex before it starts and already makes the necessary organization.
How to Transfer Chatbot Support to a Human Agent
To conclude the article, we need to understand one last point: how exactly does the transition between agents happen?
There are some tested and approved methods that we will discuss in this topic.
The important thing is to understand two main things:
- The chatbot user should have the freedom to speak with a human agent whenever they want;
- The transfer from one to the other should not be difficult or time-consuming.
Providing an adequate communication channel with the brand is even a requirement of the Consumer Protection Code.
In October 2022, some changes in the CPC started to talk specifically about support and support channels.
It was determined that the company must offer a channel that operates 24 hours a day (usually chatbot support) and another, available 8 hours a day during business hours, operated by humans.
This last one is phone support, worth mentioning. But anyway, human-operated digital support is an extra that is also worth offering.
But how to offer the transfer from digital to human support? Let’s understand some situations now, follow me:
At the Beginning of Chatbot Support
You can inform the chatbot user that human support can start at any time, depending on their request.
The message can say, for example: “Our support is automated, but you can request a human agent at any time by saying REQUEST AGENT.”
Thus, you leave it up to the users to choose the best moment to make this transition.
This is how most chatbot support happens today.
As an advantage, we can highlight the user’s autonomy, who decides for themselves which type of support is most appropriate for the situation.
And at the same time, there is also the advantage of not overloading your support team, which waits for users’ requests.
Offered by Keyword
This case is common when you do not have chatbot flows for every type of support request.
And also for cases where the request goes beyond the common support situations in your company.
In these cases, as soon as an unidentified keyword is provided to the chatbot, it offers human support.
Something like “I couldn’t understand your request, would you like to speak with a human agent?”
It is important that the user, in these cases, can choose between making their request again or directly calling the human agent.
In Special Conditions
You can also determine situations where support should always be provided by human agents, completely eliminating chatbot support in these cases.
This happens in cases like negotiation, which we exemplified above.
In this case, chatbot support ends when a specific keyword is detected.
An example: the user says “I would like to cancel my service.” The support chatbot recognizes the word “cancellation” and immediately directs the support to a human agent.
So, is everything 100% understood? I hope so!
If you have any questions, do not hesitate to comment. We answer all your questions.
You can try out our lead generation chatbot today! Just click on this link and you’ll get a 14 day trial, no credit card required.
