Exploring the Future Potential of AI-Enabled Smartphone Processors
Trailblazer community - Einstein for Service.pdf
1. Einstein for Service
Reduce service costs by driving agent
productivity at scale with AI
Kamaly Sammandamourthy, Solution Engineer
Gerald Ehret Franck, Solution Engineer
3. Agenda
01. Context and Salesforce value proposition
02. Einstein for Service
03. Demo
04. Customer stories
05. Quizz
4. The Need for Efficient Service has Accelerated
Service
Costs Customer
loyalty
Economic
Uncertainty
5. Customers Want Quality
Service, FAST
*Salesforce, State of the Connected Customer, May 2022
83%
of customers expect to interact
with someone immediately*
of customers expect companies
to understand their unique needs
& expectations*
79%
6. The Importance of Intelligent Service has Increased
Tomorrow
Today
Optimized Service, Faster Resolutions
1
Empowered Agents, Consistent Service
2
Growth Engine, Personalized Engagement
3
Fragmented, Transactional Service
2
Cost Center, Declining Loyalty
3
Growing Case Numbers & Complexity
1
79%
expect companies
to understand
their unique needs
& expectations*
*2022 Salesforce State of Service Report
7. Source: 2022 Salesforce Customer Success Metrics Survey
Einstein for Service Cloud
Intelligent service that your customers love - at scale
27% support cost
reduction
Drive Agent Efficiencies
Empower agents to do more with less with relevant articles and
recommendations
Agent Console • Self-Service• Chat Bots for SMS, Chat, WhatsApp, Facebook Messenger
Personalize Every Engagement
Unify the customer journey so agents can engage empathetically - at
scale
Agent Console • Next Best Action • Recommendation Builder • CRM Analytics
Scale Service Processes
Automate simple processes & tasks so teams can focus on the
customer experience
Agent Console • Case Classification, Routing, & Wrap Up • Article & Reply Recommendations •
Conversation Mining** In Pilot
+175B Predictions
processed daily
8. Increase Customer
Satisfaction
Service & Support
Cost Savings
Increase Agent
Productivity
Intelligent Service Delivers Better Experiences and
Faster Resolutions
32%
27%
increase in customer
satisfaction
decrease in service &
support costs
9. Let Humans Do Human Things
Automate simple
tasks to control costs
and do more with
less
Supercharge agent
productivity with
real-time, contextual
AI assistance
Empower
agents to provide
personalized service
that builds loyalty
Scale Agent Customer
10. Only Service Cloud Spans Digital Service, Contact
Centers, and Field Service in One Solution
High-Touch
Low-Touch
No-Touch
Device & Channel CDP Incidents Knowledge Cases Privacy & Preferences Communities Recommendations
& NBA
Events, Monitoring
& 3rd Party Apps
Assets
Chatbots
Messaging & Chat
Self-Service Portals
& Help Centers
AI and Automation
Voice
Visual Remote
Assistant
Field Service
Mobile App
11. Drive Efficiency with Built-in AI
Triage cases faster with real-time, intelligent recommendations
Maximize Every Interaction
Combine business rules and AI to
surface personalized offers and
next best actions in real time
Einstein Next Best Action
Quickly Triage with Accuracy
Automatically classify incoming
cases and route them to the right
queue faster
Einstein Case Classification & Routing
Find Answers Fast
Surface the most relevant
knowledge articles based on the
context of the case
Einstein Article Recommendations
Speed Up Conversation Responses
Empower agents with pre-populated
responses even as the context of the
conversation changes
Einstein Reply Recommendations
>100B
Einstein
predictions
per day
Salesforce Platform (September 2021)
13. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
+175B Predictions
processed daily
Boost Productivity for Agents and Admins
Einstein Case Classification & Case Routing
Intelligently triage cases
Classify cases by predicting relevant case details in any
language with machine learning
Automatically route work in real-time
Use your existing routing rules to deliver cases from any
channel to the right queue
Accelerate time-to-value
Intuitive, point and click setup to build predictive model
and deploy intelligence in less than a day
14. How Einstein Case Classification Works
High case
volume
Standard
and
custom
fields
Ingest feedback &
push predictions
Build the model
monthly
Populate
recommended fields
Requirements
Minimum of 1,000 cases
in the last 6 months
Improved performance with
10,000 cases
Non-empty subject & description
Available in Lightning & Classic
All languages supported
ISO, SOC2 & HIPAA compliant
Not FedRamp compliant
15. How Einstein Case Classification &
Case Routing Works
Customer
creates a
case
Einstein predicts
case fields in
real-time
Einstein verifies
Prediction
Confidence
Pre-Select Best
Prediction
Auto-Triage
and Route
(Omni)
Recommend Top 3
Predictions
Meets Threshold
Below Threshold
Above Threshold
16.
17. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
+175B Predictions
processed daily
Einstein Case Classification in Flow
Bring automagic to case triage
Build workflows based on case fields
Use Flow and machine learning to triage cases based
on AI-powered recommendations
Classify cases at the right time
Intelligently reclassify cases based on updates from
the agent
Build fast with clicks, not code
Create workflows with low code tools that make it easy
to quickly adjust and scale with changing needs
Introducing
18. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
+175B Predictions
processed daily
Close Cases Quickly and Consistently
Einstein Case Wrap-Up
Intelligently predict case field values
Leverage machine learning to suggest field values based on closed
cases and chat transcripts
Increase agent efficiency and field completion
Remove the guesswork in completing case fields and improve
reporting with accurate case details
Easily setup and deploy AI
Follow simple, guided steps to configure and build a predictive
model, then rollout to agents
19. How It Works: Einstein Case Wrap-Up
High
live case
& chat
volume
Standard
and
custom
fields
Ingest feedback &
push predictions
Build the model
monthly
Populate
recommended fields
Requirements
Minimum of 400 cases
in the last 6 months
Improved performance with
closed cases that include live
chat transcripts
Improved performance with
case subject and description
Available in Lightning
All languages supported
21. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
Decrease Time Spent Searching for Answers
Einstein Article Recommendations
Proactively review incoming cases
Leverage machine learning to review case details and text to
identify relevant articles
Intelligently surface the best article
Leverage machine learning to recommend the best article to
agents as they work in the console, based on how articles were
used on past cases
Quickly train and deploy
Select the details on cases and knowledge most relevant for
your business to train the model within hours and deploy to
agents or start with a pre-trained model
Included in UE and EE
Service Cloud licenses
22. How It Works: Einstein Article Recommendations
Knowledge
articles
Requirements
Salesforce Knowledge
100+ knowledge articles
Improved performance with
500+ knowledge attaches
Available in Lightning only
Available in English, Dutch,
French, German, Italian,
Portuguese, and Spanish
Surface recommendations
in real-time to agents
>500 article attachments on cases =
Activate pre-trained model
500+ article attachments on cases =
Build org-specific model
23. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
Einstein Article Recommendations in Flow
Build smart email auto-response workflows
Leverage machine learning and Flow to auto-respond
to cases with knowledge articles
Surface articles at case submission
Display relevant, recommended knowledge articles
during case submission
Build fast with clicks, not code
Create workflows with low code tools that make it easy
to quickly adjust and scale with changing needs
Deflect cases with automagic
Introducing
26. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
Empower Agents with Responses to Common Questions
Einstein Reply Recommendations
Build a library of approved responses
Leverage deep learning and NLP to create a catalog of
common replies based on historic chat transcripts
Surface responses to agents in real-time
Intelligently recommend the right response based on the
context of the chat conversation
Assist agents to respond and resolve faster
Enable agents to eliminate time spent typing common
responses, personalize messages and train model with
feedback
27. How It Works: Einstein Reply Recommendations
Chat
Transcripts
Requirements
Minimum of 1,000
closed chat transcripts
with 4+ chat turns per
language
Available in Lightning
only
Build recommendation
model
Build template
generation model
Surface recommendations from
template library in real-time to agents
28. How It Works: The Agent Experience
Einstein Reply Recommendations
Agent accepts
chat request
Customer sends
message
Based on the chat
conversation, Einstein
recommends the 3
most-relevant replies
Agent edits the
recommended reply and
personalizes the message
before sending to
the customer
Agent posts the
recommended
reply and the
message is sent to
the customer
Agent identifies that the
recommendation is not
helpful and feedback is
captured to improve the
model
Reply Recommendations refresh
with every customer message
29. Einstein Reply Recommendations Architecture
Included in Service Cloud Einstein Add-on
Salesforce Data Center Kleiner/EP on AWS
(US/EU/Asia/Australia)
1. Admin configures Reply
Recommendations and
triggers model training, which
in turns pulls existing Live
Chat Transcript records from
core DB to Einstein Platform
data lake.
2. Recommendation serving
model is created as part of
training which runs on
Kleiner, which uses data
pulled into Einstein Platform
Data lake.
3. When new chat starts, agents
sees recommendations in
Einstein Reply
Recommendation UI
component which uses
Serving model running on
Kleiner.
Reply
Recommendations
Setup
Chat started
Data Puller
Training/Modeling
Reply
Recommendation
Vectors
Einstein Platform
Data Lake
Quick Text
Agent Runtime UI
Serving
Reply
Recommendations
Container
Reply
Recommendations
UI component
1
2
3
31. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
Recommendations at the Point of Maximum Impact
Einstein Next Best Action
Unify your sources of insight
Combine business rules with powerful predictive
models, using Salesforce and non-Salesforce data
Surface actionable intelligence
Assist employees and customers with personalized
recommendations even as the context changes
Connect recommendations to automation
Enhance business processes with automated workflows
and triggered actions across systems
32. Bring AI-Powered Recommendations
Into Every Workflow
Improve business outcomes
Deploy real-time, personalized recommendations
to drive revenue, CSAT, and more
Build faster with clicks
Create intelligent recommendations quickly using
a point-and-click interface
Accelerate decision-making
Surface actionable recommendations by combining the
power of machine learning with Einstein Next Best Action
Introducing Einstein Recommendation Builder
33. How It Works: Einstein Next Best Action
Define
Recommendations
Create
Strategies
Surface
Recommendations
Connect to
Automation
Integrate Sources
of Insights
Salesforce Data
Partner Data
Send
Email
Load
Record
POST to
Service
Show
Screen
Collect
Data
Update
Database
Create
Marketing
Journey
Update
Case
Send
Contract
34. Einstein Next Best Action Use Cases
Upsell /
Cross-Sell
Account Info &
Data Collection
Reservation
Modifications
Shipment
Inquiries & Actions
High Bill
Inquiries
Technical
Troubleshooting
Customer
Onboarding
Applicable across many industries
35. Continuous learning loop enables better personalization
Deliver Smarter Recommendations
Products
Accounts
Order History
Recommendation
Builder (AI)
AI-Driven
Recommendations
Next Best Action
Strategy Builder
(Business Rules)
Workflow
Automation
36. Demo - Einstein Reply
Recommendations & Einstein
Next Best Action
38. Source: 2022 Salesforce Customer Success Metrics Survey
27% support cost
reduction
Optimize AI Performance
Service Analytics App
Discover insights at a glance
Easy-to-use analytics for customer and field service
Advanced Einstein metrics and take action
Prebuilt dashboards
Access insights from anywhere
Service Analytics Mobile App
40. Zenconnect Achieves Smarter Sales and
Service with Einstein AI
Challenge: Scaling and maximizing process efficiency due to rapid growth
Automating case routing and accelerating case resolution time
Increasing productivity with time being a scarce resource
Deepening insights into company data and providing predictive analytics
Solution: Einstein Case Classification predicts case field values reducing manual entry
Einstein Activity Capture detects useful information to accelerate sales cycles
Field Service Lightning connects field technicians with real-time data
Einstein Discovery aggregates customer service data to predict key figures for execs
“Einstein Case Classification
reduced time saved per ticket by
25%.”
Yohann Lecornet, CTO
3x
Increase in case fields
completed, now at 95%
25%
time saved per ticket
SERVICE SALES PLATFORM EINSTEIN
41. Hapag-Lloyd: Improved service worldwide
thanks to automation and AI
Consolidation and standardization enable customers from over 100 countries to
interact more easily
Continuous automation and process integration simplify holistic interaction with
our customers
Salesforce plays key role in customer service optimization
Faster resolution of issues worldwide as a basis for a higher-quality customer
experience
"Salesforce fits perfectly with
our quality carrier strategy to
put the customer at the center
of our processes."
Thomas Elling, Global Head of Revenue
Management and CRM
Higher
First case resolutions
>100K
Cases currently per day - with
multiple expansion stage
92%
Accuracy in case routing in interaction
with Einstein / ERP / FIS
SERVICE EINSTEIN AUTOMATION