WIXLIB

Machine Learning for Finance and RiskASUG Annual ConferenceBirgit StarmannsMay 7, 2019PUBLIC

AgendaFinance and Risk Solutions and Machine LearningFinance and Risk Scenarios Leveraging Machine LearningPredictive Scenarios Leveraging Machine LearningWrap-Up and Resources 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC2

Finance and Risk Solutionsand Machine Learning

Intelligent Technologies – drive the next-gen value economy for customers60%99%97%Of humantasks will beautomated by2025Accuracy invoice and videorecognition by2020Imagerecognitionaccuracy today 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICSource: SAP Strategy Paper – Delivering the Intelligent Enterprise, April 2018US 3.5trillionAnnual valuecreated in theenterprise4

The Intelligent Finance Functionenabled through SAP S/4HANA and SAP Leonardo agile steering and instant insights automated finance processes embedded compliance smart assistants 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC5

Vision of Finance TransformationFrom daily routine to supporting growth and new business Real-timeInsightsBackwardlookingOperations 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ ngAutomationRunningthe erInsightsSupportStrategy6

SAP S/4HANA Finance supports Your Digital TransformationEnd-to-end innovation of all your processes across your entire value-chain Better decisionswith instant, real-time insight andpredictionManufacturing& Supply Network & SpendManagement Increased performanceSAPIntelligentSuitethrough end-to-end reinventedprocesses Higher productivitywith Digital Age UX and intelligentassistanceIntelligent TechnologiesAI/ML IoT Analytics Lower TCOwith simplified architecture andcloud deploymentDigitalPlatformDataManagement 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICCloudPlatform7

Digital Technology is transforming how Business gets doneIntelligent technologies will enable finance to define its value across the enterpriseBig dataCloudReal-time insightswith in-memoryFast deploymentBusinessnetworksConnectingwith partnersUserexperienceAdapting to themodern workstyleBlockchainTransforming financeorganizationsBusinesstransactionsof the futurePredictive andcognitiveSimulating businessoutcomes 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICRobotics andmachine learningIncreased efficiency8

Technical Pre-requisites for Machine Learning On-premise SAP S/4HANA Finance Cloud: both multi-tenant and single-tenant Cloudofferings from SAP are already based on SAP S/4HANAAAABB ional Option: SAP Cloud Platform (SCP) hasCloud Extensions for SAP S/4HANA Finance Growing number of specialized appsML CashApplicationDigitalPaymentsFin StatementInsightsRealEstateSAP FioriCloudIdentity Acc.GovernanceQuickly complement new capabilitiesQuick start, easy to use, subscription-basedFlexible scalability according to changing needsSeamless integration to your existing system landscape,including on-premise, Central Finance, and Clouddeployments with replication-free, real-time integration 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICVersatile extension without disruptionRequires SAP S/4HANA FinanceSAP ERP or S/4HANA orcentral finance foundation**prerequisites dependent on solution9

Comparing approaches to automationRule enginesMachine-learningRobotic process automationEngines preloaded with highlyspecific process knowledgeenable rule-basedautomationMachine-learning identifieshidden patterns in knowledgeintensive processes andlearns from the data withoutbeing explicitlyprogrammedRobotics process automationhelps run repetitive, rulebased, and user interface–focused tasks and bridgestemporary gaps 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC10

What is Machine Learning?Sheepdog or Mop? 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC12

How does Machine Learning work?From data to tions(such as cash application)VideoSpeechPreparedataApplymodelImage and more and moreServicesCapturefeedback 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC(such as invoice processing,profile matching)13

From manual to automated processingThe quality of the recommendations of the ML service depends on the size of the training set. With theavailability of more and more historic data of user actions, the ML service gets better over time and therecommendations of the machine can be translated in automated actions.Phase 1Phase 2Phase 3 (Vision)Manual ProcessingProcessing with ML supportAutomated Processing The system logs the actions of theend users. Thereby a set of trainingdata for the ML service is created. The training data is used to trainthe ML service. The recommendations of the MLservice support the end user fortaking faster and betterdecisions. The ML service gets better withthe number of processed items. 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC As soon as the ML service isconsidered to be trained wellenough with the available historicdata, the recommendations ofthe ML service can be taken overautomatically. So more and moremanual steps can be replaced byautomated actions of themachine.This presentation and SAP‘s strategy and possible future developments are subject to change and may be changed by SAP at any time for any reason without notice. This document is providedwithout a warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, or noninfringement.14

InferenceTrainingMachine Learning Development StepsData ExplorationIterative experiments to identify MLapproach, select algorithms, determineinput parameter based application dataTraining ExecutionRegular execution of training runs, predefined model and data input, based onindividual customer dataModel PublishingValidation of model quality after trainingand activation of trained model instancesinto productive landscapesModel InferenceIntegratedConsumption 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICExposure of model as APIs providinginference based on data inputs forproductive useProvisioning of ML enabled features forend users, integrated in products andsolutions15

Finance and Risk ScenariosLeveraging Machine Learning

SAP Leonardo delivers Breakthrough TechnologiesIntelligent finance applications increase efficiency and reduce costAutomate Automating the End-to-End processesIncrease efficiency and reduce costs with SAP’sFinance Portfolio Detect and prevent fraudIdentify and rank information that positively correlateswith fraud Proactive context sensitive supportDetectPredictThe digital assistant boost productivity of your financialexperts Prediction of future valuesMore insights into the future to facilitate taking the rightdecisionsAssist 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC17

Example: SAP Cash ApplicationNext-generation intelligent invoice matching powered by machine chine LearningSAP Cash Application intelligently learns matching criteria fromyour history, reads and processes payment advice documents,and automatically clears payments with minimal intervention.AutomationSpeedReduce TCOIntegrated with S/4HANA Increase efficiency Reduce errors Enable finance to focus Learns from historical Integrated with SAPon strategic tasksFaster payment matchingReduce DSOIncrease liquidityImprove customer servicedata Learns from accountantbehavior Does not require ongoing maintenance 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC S/4HANA Cloud and OnpremiseCompliments standardrulesMaintain currentprocessing workflow18

Payment Advice ExtractionAutomated payment advice processing seamlessly integrated with SAP Cash ApplicationPaymentsAutomatic clearingComputer vision machine esDrastically lower manual efforts by automatically extracting remittance informationfrom unstructured advices (email, PDF, paper, etc.) to enhance payment matching 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ CUSTOMERPUBLIC19

SAP S/4HANA Cloud for goods and invoice receipt reconciliationEfficiently clear your open balances on the GR/IR accountsEnables intelligent processing of thereconciliation of your invoice and goodsreceipts Increases accuracy of your financial statementAccelerate period-end closing with lessmanual intervention, while ensuringcompliance with your corporate rulesEnables real-time insights into open goodsand invoice receipts for accounting andprocurement organizationsHigher efficiency through intelligentrecommendations with Machine Learning 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC20

SAP Shared Service FrameworkDeeply integrated out-of-the-box into SAP’s Finance Solution PortfolioIntegrated out-of-the-box withSAP end-to-end automation enginesClosingProcessFor example: Optimized Accounts Receivable Process withIntegrated collection and dispute worklists withSSF and the SAP Receivables Managementsolutions Simplified communication and automatedexecution of the Closing Process with SSF andSAP Closing Cockpit Enhanced issue resolution capabilities inAccounts Payable with SSF and SAP InvoiceManagement by Open Text.and many more 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICAccountsReceivableSAP SharedServiceFrameworkAccountsPayablesAnd manymore21

Predictive ScenariosLeveraging Machine Learning

Different types of predictive logicadding intelligence to finance applicationssumsumTOP-DOWN PredictionTime series algorithm on historicdata in order to predict future valuesconsidering trend, cycles and/orfluctuationPredictive Accounting on the basis ofpredicted documents being part of anindividual business process and itsdocument flowSales OrderPredictLine itemLine itemBOTTOM-UP PredictionPredictiveGoods IssuePredictiveInvoiceS/4HANA 1809 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLICThis presentation and SAP‘s strategy and possible future developments are subject to change and may be changed by SAP at any time for any reason without notice. This document is providedwithout a warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, or noninfringement.23

SAP RealSpendGain better insight into your budget and spending informationProvides visibility and access to up-tothe-minute budget and spendinginformation enabling line-of-businessmanagers perform ad hoc spendanalysis and other calculations for livebusiness processing. Integrated, real-time, and accurate view of enterprise spendSpend performance visibilityActive Spend ManagementMachine Learning for predictive spendanalysis 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC24

SAP S/4HANA Cloud for cash managementGain real-time insight into your current cash and liquidityStreamlines capital management andliquidity accounting processes byintegrating data from multiple sourcesonto a single platform. Stay on top of your company’s finances with centralized, real-time dataStreamlined Payments and BankCommunicationIn-house cash control reducing the need totransfer cash and dependency on banksReal-time insights into balances & liquidityforecast predictive analytics 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC25