Gartner s TOP 10 IT predictions By 2021, more than 50% of enterprises will be spending more per annum on bots and chatbot creations than traditional mobile app developments* - Gartner, October 2017 1 * GARTNER S TOP 10 PREDICTIONS FOR IT IN 2018 AND BEYOND
Når kunderådgivning overgår til robotter Balancen mellem kunstig- og menneskelig intelligens Kristina Grönvall, AI Strategist and Project Manager, Nordea 31.05.2018
The 4 th industrial revolution exponential change 3
What is Artificial Intelligence? Simplified definition: Machines doing tasks that would be considered intelligent if done by human SENSE COMPREHEND ACT LEARN 4
API Nova is Nordea s Virtual agent A virtual agent interacts with humans and machines to Sense Comprehend Act Support, advise and inform customers and users Understand what customers and users want Text User Interface Knowledge Base Action Voice Image Update and interact with other systems Image E-mail Avatar Action 5
Driving factors 1Improved faster customer Productivity experience gains and cost 2savings 52% does not consider it a great customer experience if they wait longer than 2 minutes 1 +55% of customers would prefer a bot to a human if the same result is achieved either way 1 By 2018, virtual employee assistants (VEAs) integrated across enterprise stacks will increase productivity of staff by 10% 2 By 2022, chatbots will be responsible for cost savings of over $8 billion annually, up from $20 million in 2017 3 6 1. LIVEPERSON 2017 2. GARTNER INSIGHT: HOW TO COLLABORATE AND COMPETE IN THE EMERGING VPA, VCA, VEA AND CHATBOT ECOSYSTEMS 3. JUNIPER RESEARCH: CHATBOTS
UNCERTAINTY DE HUMANIZATION DISCOMFORT 7 Source: Do you trust your computer
Neuroscience perspective UNDERSTANDING OURSELVES IN A LOGICAL CONTEXT Parts of the brain that supports cognition work together to read a certain interaction (images, sounds, words) Our brains are hardwired for similarity, past experience & & habit Our brains are biased for negativity & caution. This might lead to us becoming dysfunctional and disengaged To counteract the negative effects, we need to minimise danger and maximise reward 8
TRUST 9 Source: Do you trust your computer
Finding the right balance 10
What do our customers say? Need for concrete and relevant answers Need for a solution that is simple to understand and user friendly Need of overview and trust Need for personal guidance 11
Different channels for different purposes, but supporting each other 12
13 Hello my name is Nova, Nordea virtual assistant
Your brand your voice WARM FRESH Compassionate and genuine Understanding and polite Interested and helpful Open and trustworthy Lively and inspiring Able to think outside the box Passionate but down to earth Neutralizing difficult concepts CLEAR Direct and concise Clear and specific Asking for confirmation to ensure understanding EXPERT Professional and confident Making complex concepts appear simple Delegating to human if needed 14
Building content - staying conversational on the right level INSTANT AND APPEALING Conversations you could have over a cup of coffee with your neighbour DETAILS AND DEPTH Transfers for additional information such as full info about terms and conditions, customer balance history details changes, etc. 15
Transparency breeds trust Repetition Consistency Predictability Openness Technical transparency Authentication Encryption Consent, deletion of chats Information exchange Data generation Feedback 16
Our employees, our brand 17
AI should not be other it should be us Elon Musk 18
Thank you! Kristina Grönvall AI Strategy & Acceleration, Nordea kristina.gronvall@nordea.dk +45 24 66 33 90
Links Gartner: Market Insight: How to Collaborate and Compete in the Emerging VPA, VCA, VEA and Chatbot Ecosystems Gartner: Top 10 Predictions for 2018 and Beyond Juniper Research: Chatbots: https://www.juniperresearch.com/researchstore/innovation-disruption/chatbots/retail-ecommercebanking-healthcare Chatbot Magazine: Artificial Intelligence & Artificial Trust https://chatbotsmagazine.com/artificial-intelligence-artificial-trust- 5e01d32ae39a BankingTech: Putting trust in chatbots https://www.bankingtech.com/2017/01/putting-trust-in-chatbots/ LIVEPERSON 2017: https://docsend.com/view/826nkc4 PwC: Bot.Me: A revolutionary partnership: https://www.pwc.com/us/en/industry/entertainment-media/publications/consumerintelligence-series/assets/pwc-botme-booklet.pdf Business insider: Humans are willing to trust chatbots with some of their most sensitive information http://www.businessinsider.com/humans-are-willing-to-trust-chatbots-with-some-of-their-most-sensitive-information-2016-6?r=uk&ir=t&ir=t Do you trust this computer - https://www.youtube.com/watch?v=vkdqgf7lrvq 20
The key to artificial intelligence has always been the representation Jeff Hawkins, founder of Redwood Neuroscience institute 21
CONTEXT-WHY Nova uses artificial intelligence to understand the user s intent from free-text input 1Free-text «What the user types» Machine learning I need cash for a new Tesla 2 User intent «What the user wants» Confidence threshold I want to apply for a car loan Pre-defined answer 3What Nova replies Click the link below to apply for a car loan: Apply for a car loan 22