More Related Content Similar to Be Proactive in Service with Machine Learning (20) More from SAP Customer Experience (20) Be Proactive in Service with Machine Learning2. 2PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
“A breakthrough in machine
learning would be worth
ten Microsofts!”
Bill Gates (Source: AZ Quotes)
3. 3PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
• Massive improvements in
hardware(GPU, multicore)
• Deep Learning algorithms
• Big data
(e.g. business networks,
cloud applications, IoT and
S/4HANA)
• Computers learn from
data without being
explicitly programmed
• Machines can see, read,
listen, understand and
interact
What is Machine Learning? Why now?
Machines can now do things that were not possible before
4. 4PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
Intelligent Customer Service – Entire Journey
Ticket Is Created Ticket Is Processed
Communicate Resolution
Rework the Ticket
Ticket Closed
• Created based on
received information
and customer
profile.• Customer’s questions
come through
omnichannel.
• Manual or rule-based
solution assigned to
different queues or teams.
• Identify resolution
manually by searching
similar tickets and KBAs.
• Log support information.
• Feedback collection.
• Billing Triggered.
• Reply customer through
original incoming channel.
• Customer provides new
information.
• Products/Parts
recommendation
• Technician
recommendation
• Ticket classification /
prioritization
• Entity extraction
• Auto-assign ticket
• Predict Completion
Time
• Similar Tickets
• Solution
recommendation
• Knowledge graph
• OCR
• Voice-to-text
• Conversational
AI
• Recommended response
Ticket Comes in
Ticket Comes In Ticket Is Created Ticket Is Processed Work Planning Communicate Resolution
Work Planning Technician Visiting
• Capture time spent.
• Collect customer signature.
• Issue invoices.
• Assign technicians.
• Get products/parts
replacement based
on purchased
records and ticket
information.
Tickets
Insight
Tickets Insight
• Errors hotspot
detection
• Aggregated tickets
insight
Solved? No
Yes
Agent
5. 5PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
Service Ticket Intelligence enables service organizations to automate the first level
processing of incoming service enquiries, thus improving operation efficiency and satisfying
the customer’s needs.
With Service Ticket Intelligence, you can automatically categorize, assign, and answer
service tickets from various channels such as social media, e-mail, or manually created
tickets, and thus determine its workflow.
Service Ticket Intelligence uses machine learning procedures to create proposals in order to
classify tickets correctly, identify the appropriate processor, and provide a suitable solution.
The machine learning model is created and trained based on historical data of successfully
processed tickets.
Key benefits
• Natural Language Processing and Deep
learning techniques applied to understand
unstructured textand create automation
rules without explicit programming,
simplifying the implementationof automation
rules.
• Suggestedsolutions help service agents to
answer commonenquiries
Availability
• Now
Architecture
• Hosted on SAP Leonardo machine learning
platform as a business service with
communicationvia REST APIs
• Integrated with SAP Hybris Cloud for
Customeras the frontend customerservice
application
Omni-channel customer
engagementcreates
large volumes of
unstructured text data
Data in SAP customer
engagementsystems
Improve time to
resolution and
closure rate
Read ticket content,
determine category,
and automaticallyroute
ticket
Provide potential
solutions to agent
Categorize
tickets
Suggest
solution
Boost customer
experience
Machine Learning
Service Ticket Intelligence
6. 6PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
Machine Learning
Service Ticket Intelligence
• When customer sends a text
via e-mail or social post,
Machine Learning algorithms
analyze sentiments and
content
• Content groups such as
„complaints“, “service
requests“ „product support“
are identified
• Tickets are automatically
categorized and routed to the
appropriate service team
• Click for video
7. 8PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
From: XXXXXX
Subject:
Message (question):
Service category:
Incident category:
Solution provided (answer):
Typical customer support ticket attributes
Service Ticket Intelligence predicts ticket categories and solutions based on
modelling historical ticket data
Subject “message”
Category “labels”
Solution “answer”
From: XXXXXX
Subject:
Message (question):
Service category:
Incident category:
Solution provided
(answer):
‘000s of historical ticket data with associated
messages, labels and answers
Prepare data
Train model
Store model
From: XXXXXX
Subject:
Message (question):
Service category:
Incident category:
Solution provided
(answer):
New ticket arrives via
email, social platforms etc
into CRM system
Prediction Service
From: XXXXXX
Subject:
Message (question):
Service category:
Incident category:
Solution provided
(answer):
Ticket category predicted,
suggested answers for
service agent provided
“message”
“labels”
“answer”
Capture
feedback for
re-training
8. 9PUBLIC© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ
§ Ticket intelligence – auto categorization
§ Glimpse of upcoming services
Demo highlights
9. Contact information:
Kiran Karadi
Senior Director, Product Management
SAP Hybris Cloud for Customer
Kiran.karadi@sap.com
Thank you.
Contact information:
David Moore
Solution Management
SAP Hybris Cloud for Customer
david.moore01@sap.com