the age of ai conversational commerce for the next
TRANSCRIPT
IVA is the next frontier AI Powered
multi-conversation interfaces
Karl Roberts,
Head Of Propositions GCI
The Age of AI Conversational commerce for the next generation customer service
Who is GCI?
One of the UK’s leading Managed IT Service Providers.
17 years experience across 30,000 technology projects.
• Founded in 2000
• Circa £100m Revenue
• Approximately 500 staff
• 5 Core Service Pillars
• Solution / Business Outcome Focused
• Multiple Certifications Gold / Platinum
• Largest Customer 900+ Sites*
Private Cloud
Public Cloud
Hybrid Cloud
Dedicated Server Hosting
Remote/Hosted Desktop
Backup & Disaster Recovery
Microsoft Office 365
CLOUD
PCI Compliant Contact Centre
Managed Firewall
Web Filtering
Distributed Denial of Service
Authentication Services
Mobile Device Management
SECURITY AND
COMPLIANCE
Private Wide Area Network
Hybrid Wide Area Network
Rapid Connectivity
Internet Access
Local Area Network
Wireless Local Area Network
NETWORK AND
INFRASTRUCTURE
Skype for Business
Voice lines and numbers
Mobile Voice
SIP Trunks
Hosted IP Telephony
Call Routing
Mobile Gateways
SMS Services
UNIFIED
COMMUNICATIONS
Fully Managed IT Services
Flexible IT Support
IT Function Outsourcing
IT Projects
IT Review and Strategy
IT Consultancy Services
FULL IT SUPPORT
www.gcicom.net
My Background
• Oxford University : Computer Science/Software Engineering
• MIT: Machine Learning & Behavioural Physics
• 20 year Speech Recognition, 15 Years In Biometrics
• Co-Designed Infrastructure: Largest Digital Contact Center & NLU Self
Service platform in world
• Global Winner: IBM Award Innovation - Predictive Customer Services
• Global Winner: Mobile Biometrics Design
• 2000 Early AI with Procurement and then in Speech 12 years ago
• First In Europe – Passive Biometrics For Barclays Capital/Wealth
• First Cognitive Solution In Cisco Collaborative platform
• Replaced Amazons AI on Alexa, Conversational Assistant
www.gcicom.net
History – The Age Of Telephony
Why Automate?
But balance between customer service too
In the US in 1902 every 1000 telephones required 22 operators.
The Bell Telephone company employed 30,000 operators then, making
connections for 1.3 million telephones that existed. Had this ratio remained
it would have required, by 2000, over three million operators plugging and
unplugging cables to keep phones working.
About 3% of the entire US work force would be telephone operators.
Even if that many could somehow be hired and trained, the salary costs could price phone service out of reach of most people.- John Walker
www.gcicom.net
Chat Messaging Growth
Surpasses Social
www.gcicom.net
BUT Customer Service Pretty Much SUCKS
62 billion lost
due to poor
customer
service
BUT Some Success
30%
Expect More Automation
“The Rise Of The Super Agent”
Intelligent Service
Rates Expanding
80 Out of 100 Top Global
using AI Technology & 20%
year on year growth
Feb 2016
www.gcicom.net
74.5
84.8
74.0
86.4
87.5
Reduction in contact volumes
Increased company revenue/profits
Increased employee engagement
Reduction in costs
Increased customer loyalty (incl. value)
And Better CX Impacts
Customer Loyalty & Revenues
www.gcicom.net
85% can show increase
revenues/profits from improved CX
www.gcicom.net
71% Contact Centres expect
Virtual Assistant growth
79
9
210
Digital self-servicewill…
71
16
2
11
71
16
7
7
38
27
31 439
29
27
6
Digital assisted
(agent supported)
will…
Overall interactions
(spanning all
channels) will…
Headcount
employed will …
Telephone traffic
(agent led) will …
How will your contact centre develop in the next two years?71% of centres anticipate a growth in digital
agent - assisted transactions
FTE requirements will increase
However Customers
LOVE TALKING
“The Age of Conversational
Commerce”
The Future Internet – Conversational Speech, Natural
Language Understanding (NLU), Cognitive Intelligence
Mary Meeker, Perkins Caufield Byers (KCPB), 2016
Renowned Silicon Valley Analyst
www.gcicom.net
Rich Knowledge & Quality Data
Drives Higher First Contact Resolution
(FCR) Outcomes
www.gcicom.net
www.gcicom.net
CX, Deep Learning Intelligence & Knowledge
Is Going To Be Key To Virtual Assistants
www.gcicom.net
Smarter Cloud – Smarter Predictability
Reason For Enquiry?
And the…... “Next Best Action”…..Is?
But AI Predictions Are Not Perfect
But Has Bypassed Humans
www.gcicom.net
www.gcicom.net
Understand Customer Journey
Personalize & PredictChannels, Knowledge, Behaviour, Process,
Events, FCR, CSAT/NPS
IN Telephony & PBX Switch
Cloud or OnPrem
Contact Centre
Omni-Multi-Channel
PCI-GDPR Compliance
Customer[Makes Call]
PBX Switch[Connect]
IVR [Welcome Announcement]
Queuing
Music On Hold
IVR Announces[Queue PositionOr Wait Time]
Self Service [Enter CustomerReference]
Data DIP[Find Customer Details]
Agent Skill Group[Available Agent]
Transfer Call[Agent ID]
Agent Desktop [Customer Record - Screen Pop WindowOr CRM Display]
Reports &Call Logs
Wall Board [Statistics]
Agent [Checks KnowledgebaseIntranet Search]
Customer[Asks Query]
CallRecorded
IVR[Offer of CallBack Option]
Lots Of Data &
Customer
Behaviour But We
Are NOT Learning
From It
www.gcicom.net
One Conversation Model
Multiple Meta Bot Channels
Conversation
ModellerDeep Machine
Learning
Multi Mode &
Passive BiometricsDeep AI Analytics
GCI Customer Conversation Domain
IVR
Voice Mobile Web Email Social SMS IoT
Managed Service Network PCI / GDPR
Exciting Possibilities
Unlocking Customer Conversations
www.gcicom.net
Word Word Nth Word
Word Word Nth WordWord
Word
SPEAKER: Dear Sir Madam I am writing today to complain of the poor
service I received from your company on June 12 2008. I was visited by
a representative of that awful company Mr mad man at my home on that
day. Mr Mad Man was one hour late for his appointment and all for
nothing by way of apology as he arrived at noon. Your representative did
not remove his muddy shoes upon entering my house and consequently
left a trail of dirt in the hallway. Mr madman then proceeded to present a
range of products to me that I had specifically told his sister and by
telephone I was not interested in. I repeatedly tried to ask your
representative about the products that were of interest to me but he
refused to deal with my questions. We ended our meeting after 25
minutes without either of us having accomplished anything. I was most
annoyed and very upset that I wasted an entire morning and even half a
day's holiday. Waiting for Mr madman to show up. My impression of that
awful company has been tarnished and I am now concerned about how
my existing business has been managed by your firm. Furthermore Mr
mad man's inability to move his muddy shoes has meant that I've had to
engage the services and incur the expense of a professional carpet
cleaner. I trust this is not the way that awful company wishes to conduct
business with valued customers. I have been with you since the
company was founded and have never encountered such treatment
before. I would welcome the opportunity to discuss the matters further
and to learn of how you propose to prevent a similar situation from
occurring. I look forward to hearing from you. Yours faithfully Mr
Customer.
Avg : 72%
But Now Breaking New Barriers
Analytics and Multi-Lingual Possibilities
www.gcicom.net
Word Word Nth Word
Word Word Nth WordWord
Word
SPEAKER: Dear Sir Madam I am writing today to complain of the poor
service I received from your company on June 12 2008. I was visited by
a representative of that awful company Mr mad man at my home on that
day. Mr Mad Man was one hour late for his appointment and all for
nothing by way of apology as he arrived at noon. Your representative did
not remove his muddy shoes upon entering my house and consequently
left a trail of dirt in the hallway. Mr madman then proceeded to present a
range of products to me that I had specifically told his sister and by
telephone I was not interested in. I repeatedly tried to ask your
representative about the products that were of interest to me but he
refused to deal with my questions. We ended our meeting after 25
minutes without either of us having accomplished anything. I was most
annoyed and very upset that I wasted an entire morning and even half a
day's holiday. Waiting for Mr madman to show up. My impression of that
awful company has been tarnished and I am now concerned about how
my existing business has been managed by your firm. Furthermore Mr
mad man's inability to move his muddy shoes has meant that I've had to
engage the services and incur the expense of a professional carpet
cleaner. I trust this is not the way that awful company wishes to conduct
business with valued customers. I have been with you since the
company was founded and have never encountered such treatment
before. I would welcome the opportunity to discuss the matters further
and to learn of how you propose to prevent a similar situation from
occurring. I look forward to hearing from you. Yours faithfully Mr
Customer.
Traditional Way: 70-75% Success
Result : 99%
Improvement : +20% Avg : 93%
Accuracy
Empathize With Customer Behaviour
Provide Frictionless Handovers
www.gcicom.net
www.gcicom.net
And NEW Conversational Systems open up new possibilities
Offering pre-defined analytics and
improving contact center efficiency
VS
Surfacing hidden trends and patterns, and
decoding new business opportunities
Automation Without The Alienation
Detect
Context
Learn
Profile
www.gcicom.net
Forester> 70% gains in savings with
smooth human intervention
Build In Frictionless
Handoff To Humans
New Automation Benchmarks
Royal Bank Of Canada 2017
www.gcicom.net
Within 3 months of Pilot operation the system reached
93% Semantic Accuracy & 86% Task Completion
Pilot Results
October 24 2016 November 25, 2016
AUTHENTICATION SUCCESS
increased from 44% → 71.5%
WORD ERROR RATE
decreased from18% → 5.6%
SEMANTIC ACCURACY
increased from80.6% → 93.17%
Improvement of metrics
CONTINUES after transition from
pilot to full production operation!
New Benchmarks In Industry
Entirely Conversational IVR
Less than 1% negative feedback(sample of 438,333 calls)
60% of complaints related to issues that can readily be improved with further
tuning, i.e. “did not understand my request correctly”
59% Reduction in IVR Abandonment
compared to DTMF
www.gcicom.net
Royal
Bank of
Canada
IVA’s Will Be Everywhere
www.gcicom.net
“Local
Authorities”
Domain Model
GCI Serverless
Cloud Services
Home, Work and Anywhere
GCI Enterprise
Data & System
www.gcicom.net
Meet The Conversational Tax Assistant
Intent
Context
Entities
SensesEmotion Behaviour
Knowledge
Risk
Adapt
Sentiment
PCI GCPR Payment Service
META BOT
Multi-AI Conversational Intelligence
Stack Convergence!
www.gcicom.net
Me “Say Anything” AI Robot
Response + Knowledge
Context
Sentiment Tracker
Biometric
Emotion/Stress
Conversation Dialogue
TrackingEffortless Translation
ID and Verification
Biometric Analysis
PCI and GDPR Compliance
Meta Conversational Interface
Telephony Call Handling
Analyse Sentiment
Context and Dialogue Tracking
Fraud Risk Benchmarking
Multi-Lingual ConversionDemo : Karl Roberts
AI Conversational Intelligence In Action
www.gcicom.net