Introducing the BREDEX Chatbot and How It Promotes Customer Centricity

At BREDEX, customer focus begins with the initial contact and the touchpoints created specifically for this purpose. These are points of contact between potential customers or other stakeholders and the company or brand. In order to increase customer satisfaction, it was decided to reevaluate these touchpoints and improve them for the customer. As part of my bachelor’s thesis, I identified a rule-based chatbot as an improvement and then designed it. It offers users the opportunity to request information using natural language through predefined keywords or actions. This blog post describes the process that was gone through, with a particular focus on the UX measures that were taken.

What Customer Focus Means to BREDEX

In times of dynamic markets, companies that want to remain relevant to their customers must align their organization and all processes with customer requirements. A key requirement of this target group is the fast and uncomplicated processing of inquiries or, in the case of acute problems, an individual and rapid solution.
The measures we have decided on are also in line with this approach: Our goal is to present information about services and offers to customers in a way that is easy to understand and to offer accelerated and shortened communication channels.
This applies in particular to acquisition, which in future should be characterized by flexibility and speed. At the same time, we want to further lower the threshold for potential customers to make personal contact and avoid filling out forms as far as possible.

A Convincingly Modern Way of Making Contact

Traditionally, contact forms are used to gather information about potential customers. They usually consist of several fields where details such as name, position, company, type of inquiry, contact options, etc. must be entered. However, contact forms often appear very monotonous and static to prospective customers or clients and are also usually relatively long. As a result, contact forms have a high abandonment rate or are even ignored completely. In addition, they are often part of the company’s contact section and are not placed where potential customers can find information about the products. For users, this means that the hurdle to quickly getting in touch with a company is relatively high (cf. Payne, 2021).

The associated customer experience seems to have room for improvement, as customers today are used to receiving immediate answers to their questions or a response. The contact form may not be able to meet these expectations. The use of a chatbot is advantageous here due to its direct interaction with the user. The additional integration of a live chat with the sales department enables early, uncomplicated, and quick contact between both sides and expands the otherwise very limited scope of action of the chatbot.

The main reasons for using a chatbot are:

- It can take the initiative itself and ask questions about contact details at the right moment during the conversation. The information is processed in an interactive process.
- Another advantage is the suggestions already received about potential user concerns. In a contact form, the user would have to describe their problem themselves. This can have a positive effect on the user experience on the website.

- The transition to live chat avoids the long waiting times that occur after submitting a contact form. The supposed result is a better customer experience and higher lead generation for the company.

- Live chat also offers users an alternative for specific questions or if they are unwilling to talk to a machine. This means that pitches can be agreed and individual details discussed right from the moment of initial contact.

Bachelor Thesis as a Project Framework

The project was developed as part of a bachelor’s thesis and was created in collaboration with the University of Kaiserslautern and BREDEX. Within this thesis, the concept for a rule-based chatbot was designed and implemented with a focus on the user experience. In doing so, I examined the solution with the different requirements of a conversational user interface, a collective term for voice-based user interfaces that include chatbots, among other things, and identified a real problem for the user that the application should help solve as a focal point.
This project is also based on the “Chatbot Lawyer” developed by Joshua Browder in 2017. The aim of this chatbot is to gather sufficient arguments for appealing parking tickets (cf. Browder, 2021). To do this, it examines the relevant legal texts so that the user does not have to do so and can usually generate a legally binding solution for the user. The result is an extremely positive user experience based on the end result of the interaction. Another example of an effective chatbot is the Eno chatbot used by Capital. It provides customers with fast account service, notifies them of large payments, and offers a simple and quick problem reporting function. It impresses with its contextual understanding of customer inquiries and its helpful and professional manner.

How Was It Implemented?

The implementation focused on the identified user problem and our goal of guiding the customer to their destination with the help of the chatbot application. To achieve this, the concern/software problem is first identified by asking appropriate questions, and then a comprehensible and suitable solution is provided in the form of a service.
Personas and customer journeys were used in advance to provide a comprehensive analysis of the customer. Using their main characteristics and a constellation of management, sales, and marketing, a framework of relevant questions and information for the customer was created in a full-day workshop. This includes decisions and questions addressed to the user in order to best understand their concern. Methods such as brainstorming and voting were used to generate and organize ideas. With the help of the marketing department, customer-oriented formulations were then created and services were described in a compact manner. The next step was to integrate the pre-written text modules into the framework and find technical solutions to make the whole thing work. The developers involved worked with Google Dialogflow, while the data protection officers took responsibility for secure and compliant data collection. The process was rounded off by various UX measures from the design team.

Optimal UX for the Interface

When implementing CUI (Conversational User Interface), there are individual factors that require additional consideration compared to graphical interfaces. Even though current dialogues with systems include images, videos, and text elements, the interface is a blank canvas compared to the classic GUI (Graphical User Interface). As a result, the content and services usually remain hidden from the user, and interaction is based much more heavily on the user’s input. At this point, user experience design must convey to the user what they can expect within the service and provide an adequate interpretation of the feedback (cf. Følstad & Brandtzæg, 2017). By analyzing existing design checkpoints for CUI from the literature, these special features were already identified in the design phase and specifically tested in the test instances. For example, particular attention was paid to the visibility of the system status. This heuristic involves clear feedback from the system to the user as to whether the input or the goal has been understood.
Orientation aids and the identity of the chatbot are further aspects that were evaluated with great care. Through interim usability tests and a final user test with a questionnaire, design decisions were examined and the user experience was put to the test. The “usability testing” method is an excellent method due to its “hands-on” approach, which enables users to perform their typical work tasks on the system in a test scenario. It allows us to study processes on prototypes and make improvements to the user experience through feedback and observations. In the case of a chatbot, this often involves optimizing wording or navigation elements. The final test instance, in combination with a questionnaire, ensured that content, wording, and operating elements are accessible to a broad audience.

Summary, Future Prospects, and Lessons Learned

This article presented a specific measure, namely the BX chatbot. This CUI is designed to improve customer satisfaction on the website and in customer acquisition. To this end, the chatbot offers users various options for their software problem and matches these with the services offered. A future project under consideration is the use of CUI for user surveys. In this way, user-centered processes could be initiated by means of a questionnaire. This would be made available to users via the chatbot and, based on questions about user experiences and supplementary qualitative questions, would enable an evaluation of the usability and user experience of the website and other processes. Studies already conducted by Ireno Celino and Gloria Re Calegari confirm that CUIs are preferred by users when answering questionnaires. The questionnaire data is convincing in terms of reliability and response quality and can therefore be an alternative to conventional tools (cf. Celino & Calegari, 2020).

During the implementation and evaluation of usability tests and user tests, it became apparent how much influence the wording and interpretation of texts have when using a chatbot. This led to a great deal of effort being put into correcting and simplifying the texts. UX writing is one of the design trends for 2022 – and not without reason: As with UX design, UX writing also focuses on the human being in order to make the product or the desired transfer of information as effective, efficient, and satisfying as possible. With the increasing number of voice-based interfaces, it could therefore be worthwhile to establish this discipline more firmly in the future, thus turning the sober reading of information or tasks into another positive user experience.

Sources

Browder, J. (2021). DoNotPay – The World’s First Robot Lawyer. Retrieved Nov. 10, 2021, from https://donotpay.com/about/

Celino, I. & Calegari, G. (2020). Submitting surveys via a conversational interface: an evaluation of user acceptance and approach effectiveness. Preprint – International Journal of Computer Studies. Retrieved from https://arxiv.org/pdf/2003.02537.pdf

Følstad, A. & Skjuve, M. (2019). Chatbots for Customer Service: User Experience and Motivation. In Proceedings of the International Conference on Conversational User Interfaces (CUI 2019). https://doi.org/10.1145/3342775.3342784

Payne, S. (2021, February 1). Conversational Marketing: Do Chatbots Convert Better Than Web Forms. Retrieved April 10, 2022, from https://boldist.co/usability/chatbots-vs-contact-forms/