Professional Documents
Culture Documents
Unit 1: What Is Intelligent ERP in The Cloud?
Unit 1: What Is Intelligent ERP in The Cloud?
Unit 1: What Is Intelligent ERP in The Cloud?
Today's systems of record have been overtaken by new systems of intelligence, which deliver
automation, contextual collaboration, and AI. IDC calls this enhanced ERP portfolio “intelligent
ERP” or “iERP”, and it is the backbone of digital transformation.
▪ New digital business models enable enhanced user experiences and innovation, unlocking the value of
information.
▪ The best-performing companies are ahead of their peers as they digitally transform.
▪ Decades old systems of record – built in a pre-digital era – are brittle, cobbled together, and hard to retrofit to
meet today’s needs.
▪ The time is long overdue to replace them with new systems of intelligence, layering in new autonomic and
predictive intelligence assets.
▪ This enhanced ERP portfolio – iERP – will run tomorrow's businesses in an increasingly digital world.
▪ Accelerate go-live & adoption ▪ Get latest innovations & best practices
▪ Flexibly scale to new needs ▪ Keep up with regulatory changes
▪ Shift from capital to operational expenditure ▪ Easy to configure & extend
▪ No infrastructure investment needed ▪ Complete data security & protection
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 4
What Is Intelligent ERP in the Cloud?
Intelligent enterprise suite: enabling next-generation business processes
Intelligent
core
SAP C/4HANA
Capabilities
Strategies Strategies
Source: IDC MarketScape Worldwide SaaS and Source: IDC MarketScape Worldwide SaaS and
Cloud-Enabled Midmarket ERP Applications Cloud-Enabled Large Enterprise ERP
Vendor Assessment (2017) Applications Vendor Assessment (2017)
Click here for full report Click here for full report
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 6
What Is Intelligent ERP in the Cloud?
Artificial intelligence and digital technologies everywhere
Rule-Based Reasoning
Machine Learning
Deep Learning
Speech-to-Text Artificial
Intelligence
Speech
(AI)
Text-to-Speech
Robotics
Autonomics Vehicles
Machine Learning
Deep Learning
Mobile Hyper- In-Memory
Natural Language Interaction (NLI) connectivity Computing
Translation Machine Vision
Speech-to-Text Artificial
Intelligence Internet Big Data Machine
Speech
(AI) of Things Learning
Text-to-Speech
Robotics
Autonomics Vehicles
Social Cloud
Source: CMI Leonardo, Al Hilwa (2018)
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8
What Is Intelligent ERP in the Cloud?
Intelligence in SAP S/4HANA Cloud and in this openSAP course
It drives you to make better Week 1: The Future is Here: Intelligent ERP
decisions by using SAP CoPilot ▪ Intelligent ERP definition
▪ Machine learning architecture
▪ Automation and SAP CoPilot user experience
It gives you business insights Week 2: Use Cases in Finance and Procurement
by using Predictive Analytics ▪ Use case examples
▪ System demonstrations
▪ Predictive analytics configuration
It helps you automate processes Week 3: Use Cases in Sales and Manufacturing
by using Machine Learning ▪ More examples and demonstrations
▪ Implementation steps for SAP S/4HANA Cloud
▪ Future outlook on machine learning
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 10
What Is Intelligent ERP in the Cloud?
SAP S/4HANA deployment options
Cloud ERP
SAP S/4HANA Cloud SAP S/4HANA Cloud SAP S/4HANA, SAP S/4HANA,
Multi-tenant Single-tenant private managed cloud on-premise
Managed by SAP Managed by customers
▪ Configurable standardized processes ▪ Configurable and customizable
and best practices for finance, HR, processes for finance, HR, ▪ On-premise functional scope ▪ On-premise functional scope
procurement and sales procurement, sales, and other LoBs ▪ All industries supported ▪ All industries supported
▪ Defined extensibility options and ▪ SAP S/4HANA extensibility ▪ Customize, modify, and extend ▪ Customize, modify, and extend
extensions via SCP framework and extensions via SCP
▪ Annual release-cycle, speed of adoption ▪ Annual release-cycle, speed of
▪ In-depth support for professional ▪ All industries supported on customers’ schedule adoption on customers’ schedule
services and component ▪ Half-year release- cycle, customers ▪ Best-in-class deployment/upgrade ▪ Best-in-class deployment/upgrade
manufacturing; more industries on choose upgrade window within 12 efforts for on-premise ERP efforts for on-premise ERP
roadmap months
▪ BYOL or subscription ▪ Perpetual licensing
▪ Quarterly release cycle ▪ Low TCO and fast time-to-value
▪ Customer-specific system landscape ▪ Runs on customer preferred
▪ Lowest TCO and fastest time-to-value ▪ Subscription licensing
▪ Runs on SAP HANA Enterprise Cloud infrastructure
▪ Subscription licensing ▪ Dedicated system landscape on
▪ SAP-managed services on private cloud
▪ Public cloud infrastructure cloud infrastructure
infrastructure (SAP, AWS, or Microsoft
Azure)
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 11
What Is Intelligent ERP in the Cloud?
More information
open@sap.com
Follow all of SAP
www.sap.com/contactsap
NIKEiD
AUGMENTED BY AI
HUMANS –
TASKS
Human high-value
complex tasks and
knowledge work
REPETITIVE TASKS
HUMANS –
ENTERPRISE
Augmented by AI in
SYSTEM
the enterprise system
to assist in core
activities
Source: SAP CSG analysis, McKinsey Quarterly Report July 2016, Google PR, Microsoft PR, SAP Market Model
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 4
Why, and Why Now?
Intelligent ERP: core ingredients
Smart Assistants
across the ERP Suite
User experience
Conversational UX
Delivered*
Process Automation Cash application, line item matching
Lockbox
$18B
enterprise machine learning
market by 2020
94%
of companies see ML as critical
for competitive advantage!
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC Source: SAP CSG analysis, McKinsey Quarterly Report July 2016, Google PR, Microsoft PR, SAP Market Model 10
Thank you.
Contact information:
open@sap.com
Follow all of SAP
www.sap.com/contactsap
Machines driving and flying Diese Autos fahren bereits selbst und so funktioniert es Auto Bild
Dubai tests drone taxi service BBC News
Boeing invests in company building autonomous aircraft TechCrunch
Machines outperforming humans Google’s AlphaGo Zero destroys humans all on its own Cnet
in playing games AI learns to re-create Super Mario Bros. by watching someone else play it
The Verge
Robots walking and working like Atlas the next generation Boston Dynamics on Youtube
humans
Machines outperforming medical Dermatologist-level classification of skin cancer with deep neural networks
experts Nature
Machine Learning Alogitm Outperforms cardiologists Reading EKGs
HealthIT ANalytics
Why now?
✓ Big Data (for example, business networks, cloud
applications, the Internet of Things, and SAP
S/4HANA)
✓ Massive improvements in hardware
(graphics processing unit [GPU] and multicore)
✓ Deep learning algorithms
Accuracy
Testing Data Model Estimate
𝑥, 𝑦
Deployment
𝑥2
𝑥1
It’s supervised
because we used
annotated examples
to learn – rows and
columns are marked
by bounding boxes
Supervised 𝑥2
Learning
Machine Learning
𝑥1
Effects / Reward
Supervised
Learning
Machine Learning
Unsupervised Action
Learning Environment
Semi-supervised
Learning
Observation/Learning
Reinforcement
Learning A computer program interacts with a dynamic environment in which it must
perform a certain goal (such as getting the cheese out of a labyrinth),
without a teacher explicitly telling it whether it has come close to its goal.
Another example is learning to play a game by playing against an opponent.
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 12
Machine Learning Introduction
Classification of ML based on OUTPUT
𝑥2
Machine Learning
these classes (supervised learning).
𝑥1
Third-party data providers collect sales data for consumer product manufacturers to track their
market share.
The manufacturer’s hierarchy differs from the data provider’s. Given the free-text description of the
product, we attempt to classify it into the manufacturer’s six-level product hierarchy (bold fields
above).
𝑥2
Classification
Machine Learning
The outputs are continuous rather Regression
than discrete (supervised learning).
𝑥1
Time-series forecasting is well known, but many customers, such as retailers, need to be able to
interpret the forecasts.
For example, they’d like to know how much of their sales are due to uplift from a promotion,
seasonal effects, cannibalization from similar products, etc.
The model automatically learns this decomposition, as long as the relevant data (e.g., whether a
promotion was active) is provided in the historical data.
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 16
Machine Learning Introduction
Classification of ML based on OUTPUT
Classification
Machine Learning
Regression
Clustering
Density Estimation
Machine Learning
data. Regression
Unsupervised
Learning
Function
Clustering
Semi-supervised
Learning
y = h(X) Density Estimation
Reinforcement
Learning Dimensionality
Reduction
Output vector Input matrix
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 18
Machine Learning Introduction
A recipe
Yes No No
Do you need Can you write the Can you Refine your
Start
automation? rules? formulate your problem
No Yes problem
clearly?
No
Yes Do you have Yes Do you have Yes Can you Yes
Do you have Machine
a regular meaningful measure
data? learning
pattern features? success
No No No
open@sap.com
Follow all of SAP
www.sap.com/contactsap
Train
model Applications
(such as cash application)
Text Video
Prepare Apply
Speech data model
Image
…and more
… and more Services
Capture (such as invoice processing,
feedback profile matching)
SAP Leonardo
Technologies Data Intelligence Analytics
Amazon Web
Multi-Cloud Infrastructure SAP Google Cloud Platform Microsoft Azure
Services
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 5
Machine Learning Architecture
SAP Leonardo ready-to-use services
Tabular Image & Video Text Speech & Audio Business Services
General ▪ Time series change ▪ Image classification ▪ Topic detection ▪ Ticket Intelligence –
availability point detection ▪ Customizable image Classification
▪ Similarity scoring classification ▪ Ticket Intelligence –
▪ Image feature extraction Recommendation
Alpha ▪ Multi-dimensional time ▪ Product image classification ▪ Machine translation
series forecasting ▪ OCR ▪ Language detection
▪ Univariate time series ▪ Product text
forecasting classification
▪ Text feature
extraction
Roadmap ▪ Customizable ▪ Customizable image object ▪ Customizable text ▪ Speech-to-text ▪ CV Matching
recommender detection classification ▪ Text-to-speech ▪ Customer Retention
▪ Multi-dimensional data ▪ Customizable image ▪ Sentiment analysis ▪ Brand Impact
clustering segmentation ▪ Named entity ▪ Intelligent Financing API
▪ Generic classification ▪ Face detection recognition ▪ Accounts Payable
▪ Generic CSV record ▪ Face similarity ▪ Sequence to mask
matcher ▪ Human detection (S2M)
▪ Table extraction and analysis
▪ Visual document language
classifier
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 6
Machine Learning Architecture
SAP Leonardo engagement approach
Design
Explore Discover + Deliver
Prototype
High
8 11
15 5 12
17
14
Value to Company
2
16
13
4 6
10
3
7
9
Identify top 1-2 use cases Onsite user Hands-on workshop, Create clickable design prototype in Develop proof of concept of customer use
fit / gap analysis to Leonardo research show-as-is process, SAP Build case
offerings to collect deep create personas, Use SAP Fiori UX where applicable. Validated proof of concept with demo data
insights from end storyboard, create Validate prototype iteratively with by end users + involved participants
users end users via SAP Build (e.g. designers) – approx. 1 hour remote.
Ideas for to be
process, feedback in Value assessment
SAP Build, create
information architecture
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 7
Machine Learning Architecture
Example: SAP Design AppHaus and machine learning for demand forecasting with SAP Leonardo
Input Output
Train
model
Sales data
Predict
Promotions future
Prepare Apply
Weather data model
sales
Sports events
Capture
feedback
Training & Testing Data
Machine
Data Science
Learning
Analysis
Models
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8
Machine Learning Architecture
Demo
Website sap.com/ML
Enterprise ML in a Nutshell
https://open.sap.com/courses/ml1
Co-innovate ml@sap.com
open@sap.com
Follow all of SAP
www.sap.com/contactsap
RPA automates repetitive tasks done by humans without modifying business systems
Automated via
1 Automated via UI Technology 2 Connectors 2 Automate via Connectors
RPA toolsets also have the ability to use
Legacy ERP LoB Applications 3rd Party dedicated connectors beyond UI parsing,
Business e.g.
Typically multiple ▪ Database queries/SQL/stored procedures
Application
instances across MS File
(Cloud/OP) ▪ Active Directory commands
business units or Office System
locations ▪ E-mail services
▪ MS Office APIs
▪ Windows file commands, Unix shell
E-mail PDF ▪ FTP
▪ PDF extractors and OCR
▪ SOAP and REST access to Web services
Data- Web ▪ Web browser APIs
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC base Services 3
Process Automation
Typical customer issues
Requirements Features
▪ Ease of design and implementation Robot ▪ Automation methods:
▪ Integration flexibility Designer API interfaces
▪ Built-in ML capabilities OCR/NLP
Visual, UI capture
▪ Process mining & discovery
▪ Robot templates
Exception handling
RPA to ML
Requirements Features
▪ Support large-scale digital workforce Platform ▪ Security, credential management
deployment from the cloud platform ▪ Provisioning ▪ Logging and reporting
▪ Ensure stability during cloud system ▪ Orchestration ▪ Predefined robot templates
update (bot lifecycle management) ▪ Integration for business processes
Area Example
SAP Intelligent ERP ▪ Invoice posting
Core Finance
Powered by SAP HANA and RPA automation ▪ Revenue recognition
component on SAP Cloud Platform ▪ Cost review
▪ Accountancy
▪ Operational matching
▪ Financial matching
Automation Scenarios Procure-to-Pay ▪ Purchasing
Best Practices Content ▪ Business partner profile/credential matching
for SAP S/4HANA Cloud ▪ Item routing
Core Finance ▪ Payment processing
Streamlined Procure-to-Pay ▪ Quoting
Accelerated Plan-to-Product Quote-to-Cash
▪ Order entry/posting
Optimized Order-to-Cash
Project and Managed Services ▪ Pricing
Enhanced Request to Service ▪ Billing
SAP S/4HANA CLOUD System Migration ▪ Data collection
▪ Data extraction
▪ System testing
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC
▪ Transaction loading 6
Process Automation
Key considerations II
Requirements Features
▪ Attended or unattended mode Runtime ▪ Robot self-learning and via
▪ Handle exception exception handling and feedback
▪ Provide supervised learning opportunities ▪ Understand intent
▪ Interact with users in real time
Collaborate
Attended
with user
Robot
Design
Retire Develop
Execute
Upgrade Maintain
Unattended
Robot
Auto-triggered
workflows
Run Test
Robot at Runtime
Intelligent core for IPA: ready to deploy Make it customer-specific with SAP
▪ Enable high degree of automation Leonardo Services
combining RPA+ML+conversational AI ▪ Automate repeatable services/
▪ Stable robots with a combination of applications
APIs and UI automation ▪ SAP Leonardo Services embedded in
▪ SAP S/4HANA as core, orchestrate Self-learning IPA offering
heterogeneous systems
enterprise ▪ AI-driven self-healing robots for
systems uninterrupted process execution
intelligent enterprise
powered by SAP
© 2018 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8
Thank you.
Contact information:
open@sap.com
Follow all of SAP
www.sap.com/contactsap
The system is proactive, i.e. it looks at the data and suggests or automates what to do next, and
can explain why it is making each recommendation, or why it is making a given prediction, i.e.
▪ Automates routine work
▪ Makes suggestions for the user
The user can have a conversation with it using natural language, as if talking to a colleague,
and it understands what the user wants, provides answers to questions, offers suggestions for
further relevant actions, and carries out the user’s requests.
Amount achieved
A “situation” captures a current business situation, its circumstances, and possible actions
Situation
Related
Situation Condition Actions
Information
Handling
Talking to SAP CoPilot is simple and intuitive. SAP CoPilot is where the user is, the first true digital assistant and
bot integration hub for the enterprise. SAP CoPilot empowers you to get things done easier, quicker, and more delightfully.
Key capabilities
▪ Conversational user experience: Interact with your SAP systems Demos in this course
using natural language.
▪ Multi-channel: Use SAP CoPilot functionality in SAP systems, Right now…
mobile (iOS), and in channels like Slack (Beta).
▪ Built-in collaboration: Chat with colleagues, share notes and Unit 2.2 Goods Receipt and Invoice
screenshots, and create whole business objects on-the-fly with pre- Receipt shows some of these
populated content based on your conversations. collaboration features
▪ Skill Builder: Create new skills with the Skill Builder, a code-free
development tool to quickly enable conversational interfaces for
enterprise applications based on OData services.
▪ Bot Integration Hub: code-free integration of bot functionality built
with SAP Conversational AI, formerly Recast.AI (with limited
authorization capabilities)
▪ Out-of-the-box skills for SAP S/4HANA Cloud, SAP S/4HANA,
and SAP SuccessFactors (Beta)
SAP CoPilot product page SAP CoPilot road map SAP CoPilot in SAP Help Portal
Situations in SAP Help Portal Situations introduction video in SAP Help Portal
open@sap.com
Follow all of SAP
www.sap.com/contactsap