Developing AI-powered conversational interfaces

Transcription

1 Developing AI-powered conversational interfaces Sophie Nachman May DTW Nice Standards Project Director, Orange 2018 TM Forum 1

2 Orange Group : facts and figures 273 million of customers worldwide 41 billion in employees - 31 Dec Orange is one of the world s leading telecommunications operators with sales of 41 billion euros in 2017 and 152,000 employees worldwide at 31 December 2017, including 93,000 employees in France. Present in 29 countries, the Group has a total customer base of 273 million customers worldwide at 31 December 2017, including 211 million mobile customers and 20 million fixed broadband customers. Orange is also a leading provider of global IT and telecommunication services to multinational companies, under the brand Orange Business Services. In March 2015, the Group presented its new strategic plan Essentials2020 which places customer experience at the heart of its strategy with the aim of allowing them to benefit fully from the digital universe and the power of its new generation networks. Orange is listed on Euronext Paris (symbol ORA) and on the New York Stock Exchange (symbol ORAN). For more information on the internet and on your mobile: or to follow us on Orange is listed on the NYSE Euronext Paris (symbol ORA) and on the New York Stock Exchange (symbol ORAN). For more information on the internet and on your mobile: livetv.orange.com or to follow us @orangebusiness 2018 TM Forum 2

3 Orange :Essentials 2020, strategic ambition Towards an unmatched customer experience 2018 TM Forum 3

4 AI powered conversational interfaces : Chatbot architecture AI empowers NLU and Speech to text 2018 TM Forum 4

5 Orange digital guidelines for conversational User Interfaces - UI 2018 TM Forum 5

6 Research project Catalyst in open innovation mode TM Forum delivering enhanced customer experience through intelligent robots 2018 TM Forum 6

7 Chatbots classification and maturity level 2018 TM Forum 7

8 Orange AI What s next Orange Business Services Enterprises 2018 TM Forum 8

9 Chatbots acting through messaging apps, through apps or web sites Virtual assistants A chatbot (also known as talkbot, IM bot, chatterbot, interactive agent or Artificial conversational entity) is a computer program or an artificial intelligence which conducts a conversation via auditory or textual methods. Such programs are often designed to convincingly simulate how a human would behave as a conversational partner. Chatbots are typically used in dialog systems for various purposes including customer care, optimizing sales, answering FAQ s, educational sessions, entertainment. From Wikipedia Mosts chatbots are accessed via messaging apps such as FB messenger, Wechat, Whats s app, Viber, or via organizations s apps and websites Or via Virtual assistants such as Amazon Alexa, Google Assistant, and soon: Orange Djingo Djingo is the intelligent assistant from Orange, controlled by voice or text. It offers an easy and intuitive way to navigate Orange TV, manage your connected home, make a call or send a SMS, or access lots of other services. Applying true AI, Djingo learns from its users to provide ever more relevant answers. OK Djingo? From orange.com web site 2018 TM Forum 9

10 Orange digital guidelines for conversational User Interfaces UI a human inside approach to AI powered conversational interfaces Principles of designing Virtual Assistants &chatbot interactions Virtual assistant and chatbot are the new media of AI applied as direct communication to Orange customers, Despite plenty of excitement it is still early days far from its full potential as the major customer touchpoint. Chatbots offer a targeted and personal engagement with customers Both for text user interfaces and for voice user interfaces Rules such as Tone of voice To ensure consistent experiences for Orange customers any device any channel Different avatars to be used for chatbots and for human messaging to clearly differentiate them Orange ethics GDPR compliance Trust Honesty to guaranty privacy of customers 2018 TM Forum 10

11 GDPR May Trust party for customers from GDPR regulation towards privacy customer dashboard PoC from Catalyst Phygital store DTW TM Forum 11

12 Thank you 2018 TM Forum 12

13 Developing Successful AI-Powered Conversational Interfaces Andy Fruhling VP, Digital Experience Portfolio Nokia Software 2018 TM Forum 13

14 End-customer expectations have changed Digital services Empowered Buy on own terms Instant gratification Personalized Contextual Omni-channel Rent vs. buy TM Forum 14

15 Digital service provider (DSP) challenges for customer care High customer effort Technology breadth The need for experts One issue can have multiple solutions Customer effort is too high - asked to do too much to fix the problem Customer care issues can be driven by a number of different network, operations or business technology domains Requires a lot of experts with very detailed technical knowledge to improve just one step in a workflow Root cause analysis is complex due to different technology domains; agents have to sort through many workflows to find cause 2018 TM Forum 15

16 Complex troubleshooting for DSPs leads to poor customer experience Customer issue My internet is slow My home Wi-Fi isn t working My video is buffering Technical environment Different services Multiple systems and silos Many vendors Remediation process Multiple causes Many workflow paths Requires technical expertise 2018 TM Forum 16

17 Chatbots are becoming an increasingly important part of the digital customer care experience 2018 TM Forum 17

18 The role of bots in customer care Decreasing customer and operator effort +70% Issue resolution Machine-learning algorithms enable bots to continuously learn and expand capability Full autonomous care Augmented care Enhanced omni-channel care Self care Assisted care Zero-touch care due to proactive issue remediation Empower human touch to solve issues faster Humanoid interface deflects calls from human Customer solves their own problems Accelerate resolution when customer calls 2018 TM Forum 18

19 Key elements of a chatbot platform for DSPs Supports multiconversational interfaces Multi-language support Machine learning required at many levels Omni-channel support 2018 TM Forum 19

20 Benefits For customers Customers can use their preferred natural language channels 24/7 Two-way conversational support over one-way static technology (e.g. Web, IVR) No need to repeat or to do too much to make the fix happen For digital service providers Improve NPS & Customer Effort Score Fully automate the highvolume low-complexity use cases Empower tier-1 agents for high-touch interactions and white-glove customer care for time-sensitive promotions 2018 TM Forum 20

21 Summar Today s digital consumers expect a positive customer care experience with their service provider across any channel in real time A key challenge in meeting this need is the ability to quickly offer the right resolution to an issue given the complexity of the technical environment and the many paths an agent can take to solve it Chatbot platforms provide an opportunity to manage high volume, low complexity issues, leaving agents free to manage more complex issues and even upsell 2018 TM Forum 21

22 2018 TM Forum 22

23 DEVELOPING SUCCESSFUL AI POWERED CONVERSATION BY DR. MARK K. SMITH 2018 TM Forum 23

24 2018 TM Forum 24

25 2018 TM Forum 25

26 2018 TM Forum 26

27 2018 TM Forum 27

28 2018 TM Forum 28

29 Valance(noun) Prosody(noun) n-grams THE BOOK OF WORDS lɪŋˈɡwɪstɪks Lemma(noun) word-for-word wort-zum-wort mot-á-mot palabra-por-palabra 逐字地, 一字不差地 ح ر فا ب ح ر ف 2018 TM Forum 29

30 THE BOOK OF WORDS 2018 TM Forum 30

31 2018 TM Forum 31

32 2018 TM Forum 32

33 2018 TM Forum 33

34 2018 TM Forum 34

35 2018 TM Forum 35

36 2018 TM Forum 36

37 2018 TM Forum 37

38 2018 TM Forum 38

39 2018 TM Forum 39

40 AI AND ML 2018 TM Forum 40

41 2018 TM Forum 41

42 PSYCHIC COMMS 2018 TM Forum 42

43 DEVELOPING SUCCESSFUL AI POWERED CONVERSATION BY DR. MARK K. SMITH 2018 TM Forum 43