Develop your chatbot concept - Part 2

In the second part we will discuss the technical aspects and the members of your Conversational AI team.

  1. Part: Definition of the problem and the use case
  2. Part: Communication and marketing of the chatbot
  3. Part: The technical requirements of the chatbot project
  4. Part: Stakeholders and costs in the chatbot project

Part 3: The technical requirements of a chatbot project

The selection of channels

Different channels can be choosen to use a chatbot. For example, the company website, messenger channels, applications such as Microsoft Teams or social media apps are available. The following questions can help you choose the communication channel:

  • Which channels do my customers use?
  • Which channels are already used for internal communication?
  • Where is a target group that I would like to open up?

Example:

Which processes must be taken into account when developing the dialogues?

Dialogs must be developed for each chatbot. How these look like depends on the following factors:

  • To which business processes does the chatbot link?
  • Which answers should the chatbot be able to give? How many answers and question options must be entered into the chatbot?
  • Which databases or company software must be integrated?
  • What happens if the chatbot does not know what to do?

Example:

It is also recommended to set up a human handover. Here, the user is forwarded to a real service employee if the chatbot cannot process a request. Such an integration is integrated as standard on the BOTfriends X conversational middleware platform.

Part 4: Stakeholders and costs in a chatbot project

A chatbot is only as good as the team that builds, maintains and trains it.

Who belongs to my Chatbot Team?

A Conversational AI team includes a product owner, developers, content managers, service agents and a decision maker.

The Product Owner

He or she is responsible for the chatbot and its success. This team member takes over the administration, monitors the KPI's and is responsible for communication and rollout.

The Service Agent

If the chatbot is unable to answer a request, the request is forwarded to a real employee using the human handover. This employee is part of the customer service and is trained to advise customers.

But it can also be an employee from the HR department, for example, if a chatbot is used to answer employees' questions about working hours and the like.

The developers

Companies can choose whether to have the chatbot developed in-house or to outsource the development to an external provider. This is especially important if corporate systems need to be connected.

If an external provider is chosen, they ideally offer a conversational middleware platform in addition to the development of the chatbot. With such a platform, for example BOTfriends X, it is possible to scale chatbot projects and adapt the chosen technologies to changes in the market without much effort.

The Content Manager

This team member plays a central role in the success of the chatbot. It is the Content Manager, also called Intent Manager, who maintains and trains the contents of the chatbot. Therefore he or she must have the necessary expertise and know how to use the technology.

The latter requires thorough instruction and training. For example, an employee who has product expertise does not necessarily know how to create intents and uterances.

The decision maker

Whether, how and when a chatbot is introduced into a company must be decided by the responsible parties. This can be the IT department or the head of customer service.

It is important that the responsibilities and tasks within the team are clearly defined and assigned.

How much will my chatbot cost me?

The amount of the costs varies from case to case, so it is not possible to give general information here. Cost items, which however should be considered in any case are

  • Development costs or the set-up costs with a service provider
  • the cost of integration into different systems
  • the license costs for the use of software such as a Conversational Middleware Platform

 

If you want to know more about the implementation of voice and chatbot projects, we can recommend the white paper "No more POC Chatbots". In it, we discuss the most common mistakes in the planning and implementation of Conversational AI projects and explain how you can avoid them.