Abstract
This paper describes the usage of conversation user interface in financial scenario. In 2017, intelligent voice interaction has attracted a lot of attention from researchers. Intelligent voice provides a natural means of interaction. In financial scenario, users have to input and click in order to interact with the graphic interface when they use mobile banking applications. Conversational user interface based on intelligent voice technology provide feasibility to simplify the interaction between user and the application. The paper lists “transfer accounts” as a research case to show the process how the conversation interface should be used in this scenario. The research is aimed to explore richer possibilities in the future.
You have full access to this open access chapter, Download conference paper PDF
Similar content being viewed by others
Keywords
- Conversation user interface (CUI)
- Intelligent voice
- Voice user interface
- Financial scenario
- Transfer accounts
1 Introduction
Conversational user interface (CUI) is a kind of interface mode based on the chat window, which is different from the graphical user interface and natural interaction mode. The conversation interface can provide the entrance for intelligent voice, but also carries pictures, controls and hypertext forms [2, 3, 5, 7]. For complex financial service scenario, the conversation interface combines both the voice user interface (VUI) and graphical user interface (GUI) for the user to promote the overall interaction flow, in order to realize the user tasks [1, 4].
2 Background
2.1 More Natural Way of Interaction
Conversational interface based on intelligent voice technology provides a natural form for people to communicate with technology [6, 19]. From CLI (Command-line interface), to complicated GUI (Graphical User Interface), people have been searching ways to understand and communicate with machine for about thirty years (Fig. 1). Human brains are fundamentally considered that the source of speech is human. It is the first time which people use their own way to interact with machine. That means conversational interface can be considered as a much more natural way than GUI [8, 10, 12].
CUI is a transition from GUI to a VUI (Fig. 1). It is at the in the form of dialog [9, 17]. The form of dialogue weakens the way of accessing and operating the main information. It would bring not only the interface change but also the process alternation.
2.2 The Feasibility
Technology and human recognition system provide the feasibility for conversational interaction. Voice interaction technology has developed rapidly. People have the requirement that the machine is able to understand themselves as human beings, or be their real assistant. At present, more than one Technology Company has announced that their voice recognition rate is up to 97% [9, 11, 15]. In the other hand, the booming Social Networking Services (SNS) make conversational interface familiar with the users than any time in the history [18].
2.3 The Improvement of Experience
For a long time, the GUI has been used as an intermediary for people to interact with software. There are a lot of interactive logic and control rules in GUI, which cost users a lot of learning costs. Compared with the graphical user interface, conversational interface based on intelligent voice technology has a unique advantage [13]. People don’t have to look for the function they want among a big list. The voice user interface (VUI) should lead the whole task to reduce operation steps. At the same time, the graphical user interface (GUI) also provides the visual advantages complementary at the key touch-point. This is definitely a huge improvement in the user experience.
3 Process
Conversation user interface (CUI) is aimed to shorten the task flow and optimize user interaction with the system [16]. CUI provides the obvious advantage experience, but at the same time the open characteristic of CUI would lead to personalization of the user flow. Rather than a fixed step of GUI era, CUI design requires the user experience designer to predict user intent and create task-driven scenarios.
Establishing a financial scenario requires integration of the financial service characteristics. The following should be paid attention:
-
Security. It includes user account security and environmental security. Account security is included in the security of the user account behavior, such as login, transact etc. Environmental security includes the judgment on environmental interference, whether it is safe for trading in this environment.
-
Professional. Unlike other Fibo as personal assistant, CUI applied in financial scenarios should be designed with strong user intent. Therefore, rigorous logic and perfect feedback mechanism are necessary.
The core element of CUI is the scenario-based dialogue, which is task-driven and focuses on what users say and how chatbot responds. Therefore, the design process could be concluded as below (Fig. 2).
3.1 Establish the Scenario
“Transfer the remittance” in mobile banking have been chosen example for exploratory conversational interaction [14]. And the intelligent system provides the conversational financial service was name “Fibo”.
The scenario provides the following conveniences for the change of CUI.
-
The instructions issued by the user is clear and simple, which could be summarized in terms of “Verb plus object”.
-
It is a High frequency use function in mobile banking.
-
The normal interaction flow is less than or equal to three.
-
The number of modifiable or edited items could be controlled.
Compared to the GUI, users interact with the interface by pointing or clicking, VUI provide a “query to answer” flow for the users. If someone (A) wants to transfer money to someone (B) from his/her account, the chat flow below (Fig. 3) shows the possibilities of the whole scenario.
The rounded corners rectangles with dotted border show where VUI would happen. The rounded corners rectangles with solid border show where GUI would be added to help. VUI is to guide the user to speak out their intent, and turn it into command in order to push the task flow. However, GUI is to help the user dealing with complicated information, which is hard to represent by VUI (Fig. 3).
3.2 Input
Utterance is about what users say. Utterances are composed by keyword commands, natural speech and slot.
Keyword Command.
It is what users intent to do, such as book a hotel, check the weather etc. In this case, the intent of user is to transfer money to other ones. However sometimes people would not speak out their intent. The people said that “I owe jack $500” might probably want to pay back by money transfer.
Natural Speech.
Such as filler words.
Slot.
Slot is the key information point extracted from the user instruction by the NLU. The NUL module understands the specific requirements of user instructions through this key information and its Value definition (Slot-Value). In this case, the slots are payee name and money amount.the slot-value are corresponds to “jack” and “$500” (Fig. 4).
However, one of the most important aspects is defining the range of what people might say. Human language is extensive and profound. In this case, the intent of “Transfer the remittance” could be presented as follow. Transfer,send,remit,pay,wire could be used to express the meaning of the transfer (Fig. 5).
The utterance showed above is aimed to extend built-in intents. Also providing examples of utterance could help to ensure a good experience. One-shot utterance including the key command and slot could lead the user to the destination. One-shot example is always used as a start (Fig. 6).
Input is about what user communicate with the GUI shown in the dialogue box. In this case, user need to confirm payee information by browsing list and select payee account list by pointing option. The dialogue box carry the view and control for the users’ eyes and hands.
3.3 Respond
CUI responds the user through text to speech (TTS) and GUI dialogue box. Designers write a good answer scripts, played through the TTS engine into speech. The response brings the user the most intuitive experience. GUI reflects the personality of the product through interactive design and visual design, while VUI is a personalized through the emotion and speech organization. To set the VUI personality in financial scenario needs to be based on in-depth user research. Through users’ persona to insight the social need. It also needs to integrate the character of financial product, such as professional, reliable, etc.
The design of CUI requires the designer to predefine several types of responses that might arise in the conversation. The whole dialogue design is much like the design of movie script, which help users to complete tasks through query and answer.
At the same time, the research pays more attention on the touch-point which GUI should be applied. The CUI carries GUI as dialog box form to make sure the whole process in immersive environment.
There are three Key nodes in “transfer the remittance” scenario. These key nodes would produce different touch-point according to the user’s account status.
-
Payee list. The payee who is not in the list means that payee information including name and account should be newly created. There is no shortcut for inputting a long string of card Numbers, unless the payee account is mapped as an email address or cellphone Numbers. If the form of the account information is mapped as contact information, the scenario should provide VUI branch flow aimed to get the information from users’ contact list (Fig. 7).
-
Payee account uniqueness. If the payee has more than one account in the payee list, the system should ask the user to choose one. If the utterance includes the feature information such as XYZ bankcard, which could be tagged by system, the step of payee account selection should be omitted. Otherwise, GUI selection control is recommended to apply for users’ easy choosing if there are more than three accounts (Fig. 8).
-
Payment account balance. If the payment account balance is insufficient, other accounts which balance is sufficient should be matched to the process. The system should ask user to confirm the alternation. If the payment account is unique, the transfer amount less than the balance should be recommended to help to continue the process (Fig. 9).
4 Design
In this case, two session processes have been selected in this scenario to carry out the high fidelity prototype design (Fig. 9). The design follows the following principle:
-
Maximize the entrance of voice
-
Enhance content presentation
-
Use clear and simple graphical interface controls, form styles to improve interaction efficiency.
The first session choose the one-shot session. The whole conversation is brief. GUI dialogue box is used in the penultimate step for confirmation for the transaction. With the development of biometric technology, users will no longer rely on graphical user interface to verify signatures in the future.
The second session choose the payee account adulteration session. The conversion is composed of wise response and simple control. Compared to the traditional GUI process, only one point for the option is involved in the whole process (Fig. 10).
The element of CUI is dialogue box. The process of different session could be considered as the combination of different boxes.
5 Conclusion
CUI connects the graphical user interface and natural user interface. By using social scene with user familiar with, through the intelligent voice technology which help users to think and do users want, with interesting dialogue the CUI build up tasks-driven scenario. CUI have succeed in shortening the process and optimizing the user experience.
The mission driven, immersive, single-threaded elements of financial service has a good match for the CUI experience. Based on the research on CUI for financial services scene provides a practice to explore the intelligent voice commercial market.
In the future, CUI may evolve or disappear with technological innovations. But it’s still worth exploring the method of a natural interaction.
References
Aprile, W., van der Helm, A.: Interactive technology design at the Delft University of Technology - a course about how to design interactive products. In: Proceedings of the E&PDE 2011, London (2011)
Archer, B.: The nature of research. J. Codes. 2, 6–13 (1995)
Arvola, M.: Interaction design qualities: theory and practice. In: Proceedings of the NordiCHI 2010, pp. 595–598. ACM Press (2010)
Klemmer, S.R.: The future of user interface design tools. In: Extended Abstracts on Human Factors in Computing Systems, CHI 2005, pp. 2134–2135. ACM (2005)
Liebenberg, L., Steventon, J., Brahman, N., Benadie, K., Minye, J., Langwane, H., et al.: Smartphone icon user interface design for non-literate trackers and its implications for an inclusive citizen science. Biol. Conserv. 208, 155–162 (2016)
Giraudet, L., Imbert, J.-P., Bérenger, M., Tremblay, S., Causse, M.: Erratum to “The neuroergonomic evaluation of human machine interface design in air traffic control using behavioral and EEG/ERP measures” [behav. brain res. 294 (2015) 246–253]. Behav. Brain Res. 300(5), 246 (2016)
Grudin, J.: The case against user interface consistency. Commun. ACM 32(10), 1164–1173 (2017)
Princen, J., Bradley, A.: Analysis/synthesis filter bank design based on time domain aliasing cancellation. IEEE Trans. Acoust. Speech Signal Process. 34(5), 1153–1161 (2003)
Al-Tamimi, A., Lewis, F.L., Abu-Khalaf, M.: Model-free Q-learning designs for discrete-time zero-sum games with application to H-infinity control. In: Control Conference, vol. 43, pp. 1668–1675. IEEE (2015)
Khaitan, S.K., Mccalley, J.D.: Design techniques and applications of cyberphysical systems: a survey. IEEE Syst. J. 9(2), 350–365 (2015)
Esselman, P.C., Allan, J.D.: Application of species distribution models and conservation planning software to the design of a reserve network for the riverine fishes of northeastern mesoamerica. Freshw. Biol. 56(1), 71–88 (2015)
Arora, S., Arora, G., Lakshminarayan, R., Brown, G., Frid-nielsen, M., Mok, C., et al.: Hierarchical drag and drop structure editor for web sites. US, US8935602 (2015)
Bianchi, A., Corbetta, J., Invernizzi, L., Fratantonio, Y., Kruegel, C., Vigna, G.: What the app is that? Deception and countermeasures in the android user interface. In: Security and Privacy, pp. 931–948. IEEE (2015)
Marcus, A.: Universal, ubiquitous, user-interface design for the disabled and elderly. HCI and User-Experience Design. HIS, pp. 47–52. Springer, London (2015). https://doi.org/10.1007/978-1-4471-6744-0_7
Bailey, M.: Interaction design: beyond human-computer interaction 2002(4), pp. 369–378 (2011)
Kolko, J.: Introduction-thoughts on interaction design. Thoughts on Interaction Design, pp. 96–97 (2010)
Cooper, A., Cronin, D., Noessel, C., Reimann, R.: About face: the essentials of interaction design. Mental Neurol. 18(1), 1 (2014)
Xin, X., Zhou, W., Li, M., Wang, H., Xu, H., Fan, Y., et al.: Innovation design in personal center interface of mobile application. In: Marcus, A., Wang, W. (eds.) International Conference of Design, User Experience, and Usability. LNCS, pp. 310–323. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-58637-3_25
Bruce, M., David, K., Bryan, R., Ian, R., Jonathan, D., Anne, M.C.: Multi-modal assessment of on-road demand of voice and manual phone calling and voice navigation entry across two embedded vehicle systems. Ergonomics 59(3), 344–367 (2016)
Acknowledgement
We thank all my fellows for their enthusiasm and hard work. We thank Zhang Ning for her knowledge. We thank Wei Liu, Xin Xin, and Di Zhu for their support. We thank ICBC D&R center for their support. The publication of this research project was supported by the Fundamental Research Funds for the Central Universities (No. 01900-310422110).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Wang, L., Liu, S. (2018). Research on Conversational User Interface in Financial Scenario. In: Marcus, A., Wang, W. (eds) Design, User Experience, and Usability: Designing Interactions. DUXU 2018. Lecture Notes in Computer Science(), vol 10919. Springer, Cham. https://doi.org/10.1007/978-3-319-91803-7_36
Download citation
DOI: https://doi.org/10.1007/978-3-319-91803-7_36
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-91802-0
Online ISBN: 978-3-319-91803-7
eBook Packages: Computer ScienceComputer Science (R0)