Conversational AI is a technology capable of having a humanlike conversation with a user. In the context of customer support, it can be used in a few different ways, but always with the goal of offering a better customer experience.
The most common use is offering 24/7 support and reducing the number of inquiries reaching the support team. We have also seen customer support teams using conversational AI to help them find information when dealing with user questions.
Thomson Reuters is a Canadian multinational media conglomerate based in Toronto, Canada. The company has more than 25 000 employees and generates 5.5 billion in annual revenue from different solutions including their tax & accounting software Domínio.
Domínio receives more than 8,000 queries per day. These queries were handled by a team of 300 call center agents on the phone and an additional 200 agents responding via text through a ticketing system. They identified the need to respond to customer queries faster and more efficiently.
Domínio identified repetitive questions that could be answered by a chatbot. This would immediately improve the customer experience by providing simple answers faster and more accurately and giving support agents more time to deal with more complex questions.
A chatbot would also scale to service an ever-increasing volume of customers without increasing customer support headcount.
This would lead to future headcount savings and a high return on investment. On top of the ROI, Domínio would transform the customer experience for the better, which has an impact on many other fronts such as sales and product retention.
After reviewing multiple conversational AI platforms, Thomson Reuters chose to go ahead with Botpress because it did not only meet the criteria, but they were well ahead of other platforms.
The Botpress studio is an easy to use graphical interface that facilitates the collaboration of subject matter experts on building a chatbot. They can tailor the chatbot to answer repetitive questions, but also get insights on what questions users are not getting any answers. Non-technical staff can make sure the content and the experience fit the needs of their users.
Developers can focus their effort on integrating Botpress with the ecosystem of software using the unmatched flexibility of the platform.
The Botpress platform is a solution that allows for developers and non-technical staff to collaborate on a project to make sure it’s a success. The expertise of everyone is required to build a strong solution .
Domínio began with a limited scope pilot whereby their chatbot, TRIA, would respond to queries regarding one application topic.
The first important action made by Domínio is that they embedded the chatbot into their own accounting and tax software application using Botpress Webchat. This enables customers to easily access the chatbot with minimal disruption.
The second meaningful action was to create a Botpress module to extract interaction information on the user’s interaction within their own software. They could then use this information to customize the user experience of the chatbot.
For example, the visual flow builder in Botpress can use the extracted information such as which version of the software the user is using, which screen and field they are on, and what error they encountered to determine what to do next.
Using the user context to control the user experience is critical to creating a great user experience because it allows chatbot developers to make the chatbot as relevant as possible to the task at hand. It is hard to impossible to use the user context in this way on other chatbot development platforms.
Domínio did extensive testing on performance before releasing the chatbot to its customers to ensure that the chatbot could easily scale to their requirements and also be highly available. This not only validated the expectations but also removed considerable delivery risk ensuring the solution delivered met the business needs.
David Antonio, from Domínio, shared with us that Domínio intends to set up “human in the loop” in the near future so that questions that the chatbot can’t answer will be escalated to support agents in real time.
Mr Antonio also shared that they plan to shortly enhance the chatbot’s abilities by integrating it with more internal systems. In the longer term, he said they intend to expand the use of chatbots beyond internal and external technical support to marketing, training and other commercial activities.