How chatbots can work with a contact centre

What is a chatbot?

What is a chatbot?

A chatbot is defined by an online dictionary as

a computer program designed to simulate conversation with human users, especially over the Internet.

Sometimes described as Artificial Intelligence (AI), these chatbots in fact use sophisticated programming to "converse" with a user on a wide variety of subjects and respond in a realistic manner to the user's questions or statements. Such chatbots require a huge amount of programmingHowever, although most people imagine these types conversation bots if asked, businesses can use many of the features of chatbots to provide a highly effective way to automate customer engagement. 

So many businesses can use a much more focused list of responses to user questions that enable the user to answer relevant answers to questions.

Chatbot Types & Uses

Based on the experience of building our own chatbot there are various stages in the process which increase in complexity

  • Basic FAQ chatbot:  simple matching of a query with an answer.  
  • Alternative phraseology:  create different ways to recognise that the query can stated.
  • Funnel / Diagnostic questions:  this is a key step to making this into a useful chatbot.  The chatbot creates a series of diagnostic questions to narrow the from a general question to a specific answer.
  • Context based answers:  the bot to be able to understand answers like “Yes” in the context of the previous question thread.
  • Handoff rules:  identify when the chatbot should pass the enquiry across to the human agent.
  • Access to client data based on credentials.
  • Conversational Chatbot


Uses of Chatbot in contact centre context

  • Out of hours servicing: potential to provide access to basic information.
  • Managing call volumes: manages initial interactions with clients and establishes the subject matter of the conversations.  Collect information for use by the agent to solve the problem.
  • Point to resources that can potentially answer the query without needing to interact with an agent.
  • Self service activities: chatbot can facilitate the completion of activities such as making sales; pay or view invoices.
  • Integrating BOT and agent: enable the chatbot to hand over to an agent and in some cases. 

The chatbot in the call centre

The most effective current chatbots are those that have a tight focus. Many of these resemble an interactive FAQ.  These chatbots are not trying to mimic being a human.

For many contact centres, the chatbot tends to be used in two main ways:

  • Out of hours cover: answer more straightforward and frequently asked queries.
  • Prequalification of support:  more sophisticated bots can perform a diagnostic function, so that the customer can troubleshoot a problem quickly themselves and they can be passed to the support team if they cannot find an answer.

Obviously, the business is committed to providing the high standard of service for their customers, but can a chatbot help to maintain service standards while helping to bring down costs.

As a result, they are now a serious option for many companies.

So how could chatbots impact customer service?

Unfortunately, robots will take away some of current call centre jobs. Where bots will make the greatest inroads will be in the types of roles that involve routine cognitive tasks:  jobs that require performing routine or repetitive tasks. Previously, it routine manual tasks that have been most at risk of replacement by machinery. Now through bots, routine office-based roles can be replaced by software. Roles that require creativity, innovation or judgement are becoming more valuable.  

One of the advantages of chatbots is in consistency of service. The bot will provide the approved answer to clients every time.  Although Natural Language Processing has made improvements it is not to possible to run a contact centre staffed entirely with bots. So, it is essential that experienced staff also need to be there to take over where a more complex solution is required.

And the reality is that for most call centres staff and bots will increasingly working together to solve problems.

The key thing here is that customers understand and appreciate the difference between what they can expect from a chatbot and what they can expect from a live person.

Why would a customer what to interact with a bot?

Though the answer might appear to be that they wouldn’t. But in reality, for many customers, they just want an answer to their question and get on with their day. At most simple level, chatbots can be very effective is taking on the initial interaction with the customer. Much of these engagements involve gathering basic information about the customer (name, credentials, security etc) and the issue that they wish to discuss. This information can be passed onto the agent to continue the conversation.  The only issue with this approach is that the company tries to pass off the chatbot as the agent.

However, more sophisticated chatbots can use keywords in the question to find an answer that best answers the customer’s query. This can be extended to include a variety alternative phrasing of the question. This means that the customer doesn’t have to try and guess which keywords the chatbot is looking for.

The other way in which The chatbot can use the information to identify the call centre agent that this best qualified to deal with their issue. They will connect the agent and the customer and pass on the basic data. In some cases, the bot can make suggestions the agent of resources that might with to access based on the customer’s responses.

The ability of a chatbot to take a customer through a series of diagnostic questions to an answer that can resolve the issue, makes these chatbots far more useful than a simple chatbot or an FAQ that will list a large number of possible alternatives for the customer to trawl through to try to troubleshoot their problem.

And if you are thinking about building a chatbot...

The biggest issue for the implementation of chatbots is resource.  Not only is the actual programming of a chatbot a challenging task requiring a very organised approach to match user questions with appropriate answers, and for most small and medium sized companies you will probably need the support of a specialist chatbot developer.

In addition, unfortunately, building the chatbot will require the company requiring the chatbot to invest a significant amount of time to creating a list of questions and answers.

My advice is that you should keep your ambitions small and focused.  Aim to get a minimum viable solution delivered first. The danger is that as you become more aware of the possibilities you continue to add more functionality to the chatbot, resulting in rising costs and no actual benefit to your customers.   

A possible first step would be to look at the most commonly asked questions that your support team currently field and see how many of these could be answered through self service. So, the first chatbot could be an interactive FAQ. This list of questions can be continuously improved and expanded by analysing the kinds of questions being asked, this list can then be gradually be expanded to include other questions. Over time the chatbot will become smarter.

Integrating the chatbot with the human agent will potentially make a large impact.  The chatbot could providing the initial engagement. The agent can monitor the conversations with several customers and be ready to intervene to take over the conversation or supplement the chatbot replies.

Connected by VOIP Communications

You’re in good company

Pharm N Case Study
Logo New V Case Study
First Data Logo 2018 N Case Study
Help Lin N Case Study
Shelter Logo N Case Study