SAP Machine Learning and Intelligent Enterprise RoadmapBert Laws, Area Product Manager Machine Learning, SAPSession ID 82247May 7 – 9, 2019

About the SpeakersSpeaker Name Bert Laws, SAP Area Product Manager, Machine Learning10 years at SAP with a background in PerformanceManagement, Financial Solutions Rapid Deployment Solutionsand emerging technologies.

Key Outcomes/Objectives1.Clear understanding of Intelligent Technology tools2.Understand Side-by-Side vs Embedded delivery models andwhat this typically means to you3.View of product roadmaps for content Machine Learning,Conversational AI, RPA, Situation Handling and otherIntelligent Technologies over the next few releases

Agenda What do we mean by the “Intelligent Enterprise”? Delivery Models Roadmaps

The business problems that SAP Customers are solvingSales MarketingOperationsFraud RiskFinance HR Churn Reduction Predictive Maintenance Fraud and Abuse Detection Cash Flow and Forecasting Customer Acquisition Load Forecasting Claim Analysis Budgeting Simulation Lead Scoring Inventory/Demand Optimization Collection and Delinquency Profitability Margin Analysis Product Recommendation Product Recommendation Credit Scoring Financial Risk Modeling Campaign Optimization Manufacturing Process Opt. Operational Risk Modeling Employee Retention Modeling Customer Segmentation Quality Management Crime Threat Succession Planning Next Best Offer/Action Yield Management Revenue and Loss Analysis21 Industries

Examples of business questions and related algorithmsWho will churn, commit fraud, orbuy next week/next month?ClassificationHow many products will acustomer buy next month/nextquarter?RegressionHow much will be the monthlyrevenue or number of churnersnext year?ForecastingWhat is the best offer orrecommended action for acustomer or internet user?RecommendationsWhat are the groups ofcustomers with similar behavioror profile?SegmentationHow are the customers andproducts related to each other?Link Analysis

SAP S/4HANA – Automated ERPIntelligent automation using Predictive Technologies and Intelligent User Experience accelerates and improves business processes S/4HANA and the Intelligent Suiteare foundational to a trulyIntelligent EnterpriseThe intelligent enterprise offers acomplete and scalable roadmapMachineLearningBlockchainLeonardo Intelligent Technologiesand the Digital Platform areavailable todayNaturalLanguageProcessingRobotic entThe journey to the intelligententerprise will be built on nextgeneration business processesIoTIntelligentcoreSAP Cloud PlatformSupply ChainAutomated IntegrationSAP Analytics CloudAnalyticsTravel & ExpenseContingent WorkersHuman Resources Management

SAP Machine Learning enables the Intelligent EnterpriseSAP DataSAPMachine LearningBusiness Outcomes 77% of the world’stransaction revenuetouches an SAP system26 industriesIncrease revenueConversationalExperienceIntelligent Apps7 lines of businessThe world’s largestbusiness fycustomersDigital PlatformEnablinginnovations

Machine Learning embedded in SAP ApplicationsSmartCampaignsFraudDetection70 new intelligent scenarios planned between now and the end of the year*DisputeOptimizationQuotaOptimizationRoot CauseAnalysis with MLIn Dev/ Planned CustomerBehavior AnalysisCustomerExperienceManufacturing& Supply ChainDigital CorePeopleEngagementNetwork & SpendManagementSales CapacityPlanningFingerprintAnalyticsDetect AbnormalLiquidity ItemsCareer PathRecommenderGuidedBuyingData Intelligence*Subject to enanceWarehouseShift PlanningImage-basedorderingDefect CodeProposalRecommendsimilar candidatesSentimentAnalysisConversational AIPrice forecastingAssetMaintenanceOptimizationSource ofSupplyAutomationIdentifyInternal TalentSourcingRecommendationEngineInternet of ThingsPredictiveEngineeringInsightsAuto lication &EnrichmentAnalyticsDetection ofnon ticFloor mesheetdetectionSmart Defaultsin MDMFlight riskEvaluationNext-generation AIenabled hiringExpense BotSupplier RiskModellingPolicy BotIntelligent Robotic Process Automation

How are Intelligent Technologies Delivered?CriteriaCapabilitiesEmbedded ML Data remains in HANA DBReal-time inference (always current and no round-trips)Large library of ML algorithms (90 ), TensorFlow integration via EMLAccessible via standard development toolsComplexity can be hidden in view and table functionsIntegrated with ABAP management and transport systemInherited privacy and security featuresModel managementModel performance monitoring and automated re-trainingStreaming (clustering and classification) Business & ML logic reside in the SAP S/4HANA Platform Simple use cases like forecasting or trending where the algorithms have lowdemand for data, RAM and CPU time Data located in SAP S/4HANA is sufficient for model training, no need forhuge external data for training Required algorithms is provided by SAP HANA ML (e.g. PAL, APL, TextAnalysis) and handled by PAI in terms of ML model lifecycle managementand ABAPSide-by-Side ML Leonardo ML provides cloud services for scalable and distributed trainingand inference (Cloud Foundry, Kubernetes) Support for open source frameworks and algorithms (TensorFlow, scikitlearn, R) Pre-trained models for certain business use cases Access via standard REST API Algorithms and pre-trained models for working with unstructured data /fuzzy data (image, audio, video, text ) Deep learning algorithms GPU support Model management Training infrastructure ML logic resides in the SCP Platform while the business logic can be basedon SAP S/4HANA or SCP Complex use cases like image recognition or natural language processingwhere among others neural networks with high demand for data, RAM andCPU/GPU time Huge volume of external data is required for model training, main focus onprocessing unstructured data Required algorithm is not provided by SAP HANA ML, e.g. third partylibraries

Machine Learning for the Intelligent Customer ExperienceProduct road map overview – Key innovationsRecent InnovationsMarketing Personalization: Product Recommendation Offer Recommendation Intelligent Scores: Channel Affinity Best Email Sending Time Contact Engagement Score / AccountEngagement Score Sentiment Engagement Score Customer Behavior: Consumer Buyingpropensity Performance: Customer Attribution Customer Journey InsightCommerce Personalized Customer Experiences Contextual Merchandizing Context-Driven ServicesQ2/2019Sales Imaging Intelligence Sales Automation (C4C) Opportunity Scoring Lead Scoring Account Insights Configure Price Quote (CPQ) Price Optimization Up-sell & Cross sell Recommendations Configurations Recommendations Intelligent Sales Execution (Datahug) Relationship Intelligence Deal Intelligence Pipeline Management Predictive Forecasting Commissions Incentive Optimization Intelligent Coaching Sales Capacity PlanningCommerce Enterprise ChatbotSales Commissions Sales Capacity PlanningService Ticket Intelligence: Jam Articles Recommendations Ticket Classification & Entity Extraction NLP Estimated Time toCompletionBeyond Q2/19:Product DirectionMarketing Customer Behavior: Lead Conversion Propensity Customer RetentionService Ticket Intelligence: Ticket Routing Completion Spam Classifier Solution Intelligence: Email Template Recommendation Responses Recommendation KB Article Recommendation Virtual Assistant „Answer Bot“ Field Service Intelligence Service & Parts RecommendationService Ticket Intelligence Self-Service digital interface Single level Ticket Categorization Similar Tickets 2019 SAP SE and an SAP affiliate company. All rights reserved PUBLICThis presentation and SAP’s strategy and possible future developments are subject to change at any time for any reason without notice. This document is provided without a warranty of any kind, either express or implied, including but not limited to the impliedwarranties of merchantability, fitness for a particular purpose or non-infringement.CustomerExperience

Machine Learning for the Digital CoreDigital CoreProduct road map overview – Key innovationsRecent InnovationsFinance Cash Application Receivables Line Item Matching Payment Advice Extraction Lockbox Line Item Matching Payables Line Item Matching SAP Tax Compliance SAP Business Integrity Screening Financial Account Reconciliation SAP Real Spend SAP Financial Statement InsightsNatural Language Processing: Manage Bank Statements Display Correspondence History Monitor Payments Manage Supplier Line Items Process Collections Worklist Manage Payment Advices Process Receivables Manage Cost Centers Approve Bank Payments Doubtful Accounts ValuationQ2/2019Sales Sales Quotation Sales Performance Delivery PerformanceNatural Language Processing: Manage Sales Contracts Manage Sales Orders Manage Sales Quotations Track Sales Orders Manage Credit Memo Request Manage Debit Memo Request Manage Sales Orders without chargeProcure Quantity Contact Consumption Cash Discount at Risk Creation of New Catalog Item Propose Material Group Reduce Off Contract Spend Predict Delivery Date for PurchaseOrderNatural Language Processing: Smart BuyingBeyond Q2/19:Product DirectionIdea Project Cost Forecasting Digital Content ProcessingNatural Language Processing: Manage ProjectsFinance Detect Abnormal LiquidityItems Intelligent AccrualrecommendationFinance Cash Application for FICAProduce Stock In Transit Demand-Driven ReplenishmentProcure Image-based OrderingIdea Variant Configuration analytics forconfiguration data – Top SellerService Automated Email and ServiceRequest Category mappingDigital Support Experience withNatural Language Processing Create Feature Request Create Support Incidents Contact Key User S-User Management Advanced Search Show My Support Incidents Transport Management with NaturalLanguage Processing Manage Freight Agreement Produce Defect Code Proposal (incl.Text Recognition)Procure Intelligent Approval WorkflowProduce Early detections of slow and non movingstocksMaster Data Management Business Rule Mining Smart Default ValuesSCP Enablement Automatic Floor Plan ExtractionService Configuration optimization based onload in Convergent Charging 2019 SAP SE and an SAP affiliate company. All rights reserved PUBLICThis presentation and SAP’s strategy and possible future developments are subject to change at any time for any reason without notice. This document is provided without a warranty of any kind, either express or implied, including but not limited to the impliedwarranties of merchantability, fitness for a particular purpose or non-infringement.

Machine Learning for Manufacturing & Supply ChainManufacturing& Supply ChainUse Case Portfolio for Digital Supply ChainRecent InnovationsIntelligent Asset ManagementPredictive Maintenance & Service Machine Failure Prediction Unnormal Machine State Detection Failure Mode Analytics Extensibility of ML Engine throughcustom algorithmsDigital ManufacturingPredictive Quality Management Defect Detection Extending Defect Detection withmachine dataResource Orchestration Auto Dispatch - Preload ResourceallocationSupply Chain Management & LogisticsIBP Demand Sensing (Leveraging patternrecognition techniques to create shortterm forecast on a daily basis) Custom Alerts with Machine Learningbased Outlier Detection New Forecasting AlgorithmQ2/2019Intelligent Asset ManagementPredictive Maintenance & Service Life Indicator Forecasting Leading Indicator AnalyticsDigital ManufacturingPredictive Quality Management Defect detection – improvement datapreprocessingResource Orchestration Auto Dispatch - PreloadResource allocation Enhanced with UserInterfaceSupply Chain Management & LogisticsIBP Anomaly detection in Batch JobsBeyond Q2/19: Product DirectionIntelligent Asset ManagementPredictive Maintenance & Service Fingerprint Analytics using anomaly detection Configuration Correlation AnalysisAsset Manager Measure recognition and equipment recognition viaobject detectionAsset Strategy and Performance Management Cost-sensitive Asset Maintenance StrategyoptimizationPlant Maintenance (PM) / Execution Failure Mode Suggestion forWorkorders/Notifications Intelligent Work order ranking Notification/Alert deduplication based on MLtechniquesSAP Newton Predictive Engineering Insights Frequency diagnostics for rotating equipmentSupply Chain Management & LogisticsIBP New Product Introduction Decision Support for Alert Handling Job scheduling optimization Anomalies Detection in Master Data Forecast Level Optimization Natural Language Processing via CoPilotExtended Warehouse Management (EWM) Intelligent Fixed Bin StrategyWarehouse Insights (WI) Warehouse Shift Planning including What-if analysis Maximum Working Capability Analysis for ResourceType Intelligent Optimization Settings for WarehousesTransportation Management (TM) Automatic Transit Time Adjustment Real Cost Tour Planning Intelligent Transportation CockpitDigital ManufacturingPredictive Quality Management: Golden batch- Anomaly detection 2019 SAP SE and an SAP affiliate company. All rights reserved PUBLICThis presentation and SAP’s strategy and possible future developments are subject to change at any time for any reason without notice. This document is provided without a warranty of any kind, either express or implied, including but not limited to the impliedwarranties of merchantability, fitness for a particular purpose or non-infringement.

Machine Learning for the Indus