Bot-based IT Troubleshooting Benjamin Braun advised by Jonas - PowerPoint PPT Presentation

Chair of Network Architectures and Services Department of Informatics Technical University of Munich Bot-based IT Troubleshooting Benjamin Braun advised by Jonas Jelten, Simon Bauer Friday 11 th January, 2019 Chair of Network Architectures and Services Department of Informatics Technical University of Munich

Chair of Network Architectures and Services Department of Informatics Technical University of Munich Agenda Value of Chatbots Current Use Cases Difficulties and Problems in IT Conclusion Bibliography B. Braun — Bot-based IT Troubleshooting 2

Value of Chatbots • Ongoing issue: Improvement of customer service B. Braun — Bot-based IT Troubleshooting 3

Value of Chatbots • Ongoing issue: Improvement of customer service • Chatbots: • Dialog system for human-computer-interaction • Emulation of human behavior in a conversation • Providing both chatter and serious help with tasks and questions • Heavy development over last decades B. Braun — Bot-based IT Troubleshooting 3

Value of Chatbots Advantages of chatbots over conventional customer services • More personal and interactive than a FAQ • Workload shift from customer to program • Possibility to ask for additional data from the user • Immediate responses • Permanent availability • Access to complete knowledge base B. Braun — Bot-based IT Troubleshooting 4

Value of Chatbots Benefits in costs • Setup + maintenance costs B. Braun — Bot-based IT Troubleshooting 5

Value of Chatbots Benefits in costs • Setup + maintenance costs • vs. salaries for customer service team B. Braun — Bot-based IT Troubleshooting 5

Value of Chatbots Benefits in costs • Setup + maintenance costs • vs. salaries for customer service team • Important: Chatbots not capable enough yet to replace humans completely. • 50% - 90% of queries are processable, percentage increasing [8] B. Braun — Bot-based IT Troubleshooting 5

Value of Chatbots Benefits in costs • Setup + maintenance costs • vs. salaries for customer service team • Important: Chatbots not capable enough yet to replace humans completely. • 50% - 90% of queries are processable, percentage increasing [8] • Reduction in workload for service agents towards only skill-challenging requests ⇒ More fulfilling experience for staff ⇒ More motivated, happier staff ⇒ More satisfied customers B. Braun — Bot-based IT Troubleshooting 5

Value of Chatbots Benefits in costs • Setup + maintenance costs • vs. salaries for customer service team • Important: Chatbots not capable enough yet to replace humans completely. • 50% - 90% of queries are processable, percentage increasing [8] • Reduction in workload for service agents towards only skill-challenging requests ⇒ More fulfilling experience for staff ⇒ More motivated, happier staff ⇒ More satisfied customers ⇒ Long term investment, but already cutting costs by at least 55% since deployment [8] B. Braun — Bot-based IT Troubleshooting 5

Current Use Cases History and development • First chatbot: ELIZA by Joseph Weizenbaum (1966) [7] • Designed to fool people, not to solve tasks • Trend for research of following decades: Trying to pass Turing test [1] B. Braun — Bot-based IT Troubleshooting 6

Current Use Cases Personal assistants • Siri, Alexa, Google Assistant, Cortana, Bixby, ... B. Braun — Bot-based IT Troubleshooting 7

Current Use Cases Personal assistants • Siri, Alexa, Google Assistant, Cortana, Bixby, ... • Many capabilities related to their integration, acting as voice commanding tool • Providing current information and looking up facts • Interacting with phone functions • Accessing media libraries • Controlling smart homes • ... [6] B. Braun — Bot-based IT Troubleshooting 7

Current Use Cases Personal assistants • Siri, Alexa, Google Assistant, Cortana, Bixby, ... • Many capabilities related to their integration, acting as voice commanding tool • Providing current information and looking up facts • Interacting with phone functions • Accessing media libraries • Controlling smart homes • ... [6] • Gaining more and more power (e.g. Google Duplex [5]) B. Braun — Bot-based IT Troubleshooting 7

Current Use Cases Shopping assistants/Troubleshooting bots • Used by a number of companies (DAX and MDAX: 12 out of 80) [4] • Variety of fields B. Braun — Bot-based IT Troubleshooting 8

Current Use Cases Shopping assistants/Troubleshooting bots • Used by a number of companies (DAX and MDAX: 12 out of 80) [4] • Variety of fields • Often very specialized, e.g.: • Adidas: sport course booking • BMW: broadcasting for tournaments • Persil: spot removal • VW: second-hand car search • Zalando: personal shopping consultant B. Braun — Bot-based IT Troubleshooting 8

Current Use Cases Shopping assistants/Troubleshooting bots capabilities • Capabilities in general: • Often presenting as visible chat partner (avatar) • Understanding and outputting natural language • Including smalltalk and chatter • Storing details about user • Access to database for answering questions and proactively showing additional content • Referring to supplementary services if necessary • Troubleshooting capability included in some: • Checking "symptoms" with search-tree until problem is identified • Proposing solution or referring to specialized help B. Braun — Bot-based IT Troubleshooting 9

Current Use Cases Examples • IKEA’s Chatbot "Anna": • Online for over 10 years until retirement in 2016 • Should guide customers around the website and answer questions • Frustrating users with inability to fulfill tasks • Tried too hard to be natural and human-like, diverted too far from actual purpose [2] Figure 1: Avatar of "Anna" 1 1 Source: http://www.andrecastro.info/a/works.html B. Braun — Bot-based IT Troubleshooting 10

Current Use Cases Examples • "SuperAgent": • Browser extension, not bound to a specific company • Leverages publicly available e-commerce data • Processing product information, user Q&A and user reviews for data and opinion mining [3] B. Braun — Bot-based IT Troubleshooting 11

Current Use Cases Examples Figure 2: SuperAgent in use on Amazon.com 2 2 Source: Cui et al. [3] B. Braun — Bot-based IT Troubleshooting 12

Difficulties and Problems in IT Utilization in IT • Barely any chatbots in IT companies • Found examples: "Assisted Search" from Dell, "Tinka" from T-Mobile B. Braun — Bot-based IT Troubleshooting 13

Difficulties and Problems in IT Special difficulties in customer service • Extensive problem causes and symptoms ⇒ Setup of databases and search trees difficult B. Braun — Bot-based IT Troubleshooting 14

Difficulties and Problems in IT Special difficulties in customer service • Extensive problem causes and symptoms ⇒ Setup of databases and search trees difficult • Broad variation in user knowledge and technical know-how ⇒ Chatbot needs to identify level of technical understand and adapt accordingly B. Braun — Bot-based IT Troubleshooting 14

Difficulties and Problems in IT Special difficulties in customer service • Extensive problem causes and symptoms ⇒ Setup of databases and search trees difficult • Broad variation in user knowledge and technical know-how ⇒ Chatbot needs to identify level of technical understand and adapt accordingly • Vague problem descriptions or wrong self-diagnosis ⇒ Advanced language processing and independent, out-of-the-box thinking necessary B. Braun — Bot-based IT Troubleshooting 14

Conclusion • Distinctive advantages of chatbots in customer services • Strongly growing capabilities of dialog systems and artificial intelligence B. Braun — Bot-based IT Troubleshooting 15

Conclusion • Distinctive advantages of chatbots in customer services • Strongly growing capabilities of dialog systems and artificial intelligence • Utilization of this technology in more and more domains • IT sector facing additional difficulties • Benefits still applicable • Greater intelligence and adaptability necessary for excellent IT troubleshooting chatbots B. Braun — Bot-based IT Troubleshooting 15

Bibliography [1] L. Bradesko and D. Mladenic. A survey of chatbot systems through a loebner prize competition, 2012. [2] P . Brandtzaeg and A. Følstad. Chatbots: changing user needs and motivations., 2018. [3] L. Cui, S. Huang, F. Wei, C. Tan, C. Duan, and M. Zhou. Superagent: a customer service chatbot for e-commerce websites, 2017. [4] J. M. Kawohl and S. Haß. Customer service 4.0 - wie gut sind chatbots? https://www.heise.de/downloads/18/2/5/4/1/3/4/2/Studie_chatbots.pdf . [5] Y. Leviathan. Google duplex: An ai system for accomplishing real-world tasks over the phone. https://ai.googleblog.com/2018/05/duplex-ai-system-for-natural-conversation.html . [6] T. Martin and D. Priest. The complete list of alexa commands so far. https://www.cnet.com/how-to/amazon-echo-the-complete-list-of-alexa-commands/ . [7] M. L. McNeal and D. Newyear. Introducing chatbots in libraries, 2013. [8] I. Steele. It’s all about the $$$ - how much money can chatbots actually save you? https://www.comm100.com/blog/how-much-can-chatbots-actually-save-you.html . B. Braun — Bot-based IT Troubleshooting 16