1、October 21-24,2024Mandalay Bay Convention CenterLas Vegas,NevadaSession 2082Jan-Peter PreuAdvisory Partner Technical Specialist for AIIBMPlanning Analytics,LLMs,and watsonx.ai better together#IBMTechXchangeAgenda&What you will learn2IBM TechXchange|2024 IBM CorporationAgenda010203040506Planning Anal
2、ytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases for Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM CorporationAgenda010203040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases fo
3、r Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchangeTM1 as heart of IBM Planning AnalyticsQ2Q1Q3Customer 3Customer 2Customer 1Product 3Product 2Product 1ProductTimeCustomerWhat is TM1?TM1(Table Manager 1)is a multidimensional,in-memory OLAP database with a cell-orient
4、ated structurelike spreadsheetsthat allows users to create sophisticated financial models and perform advanced calculations while benefiting from control and governance.Read more IBM TechXchange|2024 IBM Corporation5#IBMTechXchange6Flexible deployment optionsSuperior scale to enable AIBuilt-in repor
5、tingUse ExcelOne integrated platformEnergy dataWaste dataCO2 emissionsCRM dataPayroll dataConsumptionSupplier dataEmployee trainingCollect I Organize I Analyze I InfuseSales dataERP dataCO2 emissionsSustainabilitySustainability AnalyticsStrategic planning to Net ZeroProcurement optimizationSupply Ch
6、ain or Asset simulationForecasting of initiatives,emissions(GHG)Connection to financial perspectives(Tradeoff)OPERATIONS&HRDemand planningSupply chain planning,inventory planningSales&operations planningWorkforce planningHeadcount and staffing planningSalary&compensation planningMARKETINGPromotion p
7、lanningRevenue planningCustomer profitabilityCustomer churn analysisITIT portfolio planningProject planningIT budgetingFINANCEFinancial planning&analysisStrategic planningCapital planningExpense planningSALESSales forecastingSales territory planningSales quota planningSales capacity planningIntegrat
8、ed Continuous Planning and AnalysisStart anywhere and go everywhere with IBM Planning AnalyticsPrescriptiveIBM TechXchange|2024 IBM CorporationAgenda010203040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases for Planning AnalyticsConclu
9、sionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchangeBusiness Needs for Generative AIIBM TechXchange/2024 IBM CorporationCant he get that information this on his own?Havent I answered that a few months ago?How can I put these findings into words now?I have to build a view in Planning Analytics
10、for that first Mon 2024-09-30 11:31 Todays To-Do No.1849Agenda010203040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases for Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchangeTake your Plans to the Next L
11、evel with IBM watsonxwatsonx.dataCombine sources of structured+unstructured data in one placeAs flexible as a data lake,as scalable as a data warehouseOne query to transfer data from all sources to Planning AnalyticsCollect context information about your datawatsonx AssistantConsume watsonx.ai servi
12、ces through chat in natural languageSecurely relieve controllers from manual processes and ad-hoc queriesConvenient way to input,validate and query datawatsonx.aiLLM-powered,on-demand reporting of your numbersLet generative AI write business rules and queriesProvide tailored context information in n
13、atural language Extract Entities from unstructured dataUtilize sophisticated forecasting capabilities11IBM TechXchange|2024 IBM Corporation#IBMTechXchangePlanning AnalyticsPlanning AnalyticsIBM CloudAWSOn-PremiseTM1 Databasewatsonx.ai(incl.watsonx.ai(incl.MLOpsMLOps,SPSS,SPSS,GenAIGenAI,Decision Opt
14、imization),Decision Optimization)IBM CloudOn-PremiseDeployment Space(WML)Prompt/ML/DO ModelDeployQueryRead/WriteReadData Connections(incl.PA,wxd,SAP OData)PAW/PAfEIBM+Open-Source LLMsArchitecture-watsonx.ai and Planning AnalyticsDirect IntegrationIBM TechXchange|2024 IBM Corporation#IBMTechXchange12
15、IBM Cloud/On-premise ClusterClient Environment Webchat UI/Slack/MS Teams/Userwatsonx AssistantCustom middlewarePlanningAnalyticswatsonx.aiArchitecture-watsonx.ai and Planning AnalyticsThrough watsonx AssistantWatson Discoverywatsonx.dataIBM TechXchange|2024 IBM Corporation#IBMTechXchange13Agenda0102
16、03040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases for Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchange15On-Demand Report GenerationUse Case#1IBM TechXchange|2024 IBM CorporationProject Project Pand
17、APandA:A watsonx-supported Assistant for better handling of Planning AnalyticsChallengeChallengeThe cost center managers of the client are unsure and overwhelmed when dealing with Planning Analytics due to many other tasks.Therefore,on the one hand,user rates are low due to the complexity and lack o
18、f knowledge;on the other hand,this leads to the evaluation of incorrect or irrelevant results.Although there is an abundance of information available,there is a lack of knowledge about what can be found where.Accordingly,the cost center managers approach their colleagues in Controlling with question
19、s,either to ask them for advice or to have them take on the task of finding the desired reports,for example.Value for the Client:Value for the Client:Higher user experience and rates Time savings for cost center managers and controllers Higher productivity of cost center managers and controllersSolu
20、tionSolutionThe PandA assistant was developed inco-creation and integrated into Planning Analytics.It is based on the watsonx Assistant,Watson Discovery and watsonx.ai technologies and addresses the following Use Cases:Guided dialog:Guided dialog:Creation of new,personalized views Existing reports:E
21、xisting reports:access to existing reports,dashboards and applications Free text input:Free text input:view creation in expert dialog for those who know and can describe their information needs Knowledge database:Knowledge database:A product-specific knowledge database provides access to expert know
22、ledgeAs part of the pilot,PandA was developed in natural German language.Use Case:GenAI,Market:DACH,Industry:LogisticsFurther Highlights:Further Highlights:Implementation of watsonx in widely used and leading product Planning Analytics Successful collaboration with business partner from the beginnin
23、g Upcoming external referenceAbout the Client:About the Client:The client provides fulfillment services for e-commerce companies,i.e.processes related to the fulfillment of promised services after the contract has been concluded.This includes all steps in the shipping process chain from the time of
24、the customer order,such as warehousing,picking and shipping through to returns processing.Ein Bild,das Clipart,Grafiken,Schablone,Symbol enthlt.Automatisch generierte BeschreibungIBM TechXchange|2024 IBM Corporation#IBMTechXchange16#IBMTechXchangeGuided DialogGuided DialogExpert ModeExpert ModeUse C
25、ase#2Self-Service Data Access172Free-text Query1Dynamic Dimension Elements(Cost Centers)3LLM-generated MDX view displayed in PAW UIIBM TechXchange|2024 IBM Corporation#IBMTechXchangeExistingExisting ReportsReportsKnowledge BaseKnowledge Base1Provide collection of existing reports based on query in n
26、atural language2LLM-generated answer to users question based on internal groundings3Reference&link to sourceUse Case#3Internal Repository of Reports and Knowledge18IBM TechXchange|2024 IBM Corporation#IBMTechXchangeUse Cases#1-#3 combined19IBM TechXchange|2024 IBM CorporationSearch Existing ReportBu
27、ild own CubeviewAsk context questionsPrompt LLM to summarize findingsStore findings in database#IBMTechXchange20TM1 Code Assistant:Business RulesUse Case#4IBM TechXchange|2024 IBM Corporation#IBMTechXchange21Use Case#4IBM TechXchange|2024 IBM CorporationTM1 Code Assistant:Feeders#IBMTechXchange402 6
28、14Use Case#5Entity Extraction22IBM TechXchange|2024 IBM Corporation#IBMTechXchange23Use Case#5IBM TechXchange|2024 IBM CorporationEntity Extraction#IBMTechXchange24Advanced Predictive ForecastingUse Case#6IBM TechXchange|2024 IBM Corporationhttps:/huggingface.co/ibm-granite/granite-timeseries-ttm-v1
29、Agenda010203040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics connectivityAI Use Cases for Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchangeConclusionIBM TechXchange/2024 IBM CorporationCant he get that information this on hi
30、s own?Havent I answered that a few months ago?How can I put these findings into words now?I have to build a view in Planning Analytics for that first Mon 2024-09-30 11:31 Todays To-Do No.18426Agenda010203040506Planning Analytics OverviewBusiness Needs for Generative AIwatsonx&Planning Analytics conn
31、ectivityAI Use Cases for Planning AnalyticsConclusionQ&AIBM TechXchange|2024 IBM Corporation#IBMTechXchange28Q&AIBM TechXchange|2024 IBM CorporationJan-Peter PreuIBM,Advisory Partner Technical Specialistjan-https:/ YouJan-Peter PreuAdvisory Partner Technical Specialist,IBMjan-IBM TechXchange|2024 IB
32、M Corporation#IBMTechXchangeNotices and disclaimersCertain comments made in this presentation may be characterized as forward looking under the Private Securities Litigation Reform Act of 1995.Forward-looking statements are based on the companys current assumptions regarding future business and fina
33、ncial performance.Those statements by their nature address matters that are uncertain to different degrees and involve a number of factors that could cause actual results to differ materially.Additional information concerning these factors is contained in the Companys filings with the SEC.Copies are
34、 available from the SEC,from the IBM website,or from IBM Investor Relations.Any forward-looking statement made during this presentation speaks only as of the date on which it is made.The company assumes no obligation to update or revise any forward-looking statements except as required by law;these
35、charts and the associated remarks and comments are integrally related and are intended to be presented and understood together.2024 International Business Machines Corporation.All rights reserved.This document is distributed“as is”without any warranty,either express or implied.In no event shall IBM
36、be liable for any damage arising from the use of this information,including but not limited to,loss of data,business interruption,loss of profit or loss of opportunity.Customer examples are presented as illustrations of how those customers have used IBM products and the results they may have achieve
37、d.Actual performance,cost,savings or other results in other operating environments may vary.Workshops,sessions and associated materials may have been prepared by independent session speakers,and do not necessarily reflect the views of IBM.Not all offerings are available in every country in which IBM
38、 operates.Any statements regarding IBMs future direction,intent or product plans are subject to change or withdrawal without notice.IBM,the IBM logo,and are trademarks of International Business Machines Corporation,registered in many jurisdictions worldwide.Other product and service names might be t
39、rademarks of IBM or other companies.A current list of IBM trademarks is available on the Web at“Copyright and trademark information”at: TechXchange|2024 IBM Corporation3032Usage of the chatbot was easy to understandeasy to understand.I strongly agree/I agree8 8 of 8 TestersThe Chatbot understood my
40、inputs wellunderstood my inputs well.6 6 of 8 TestersThe answers of the chatbot were informative and helpfulinformative and helpful.7 7 of 8 TestersThe chatbot could deal well with false inputsdeal well with false inputs.4 4 of 8 TestersThe usage of the chatbot was easyeasy.8 8 of 8 TestersWould you
41、 useuse PandA in your daily job?8 8 of 8 TestersWould you recommend PandArecommend PandA?8 8 of 8 TestersWhat timetime do you estimate PandA would savesave you for a view creation and information retrieval compared to todays methods?Usability and time savingsResults of user testsTesterTesterA AB BC
42、CD DE EF FG GH H5%of usual time effort5%of usual time effortXXX10%of usual time effort10%of usual time effortXXX20%of usual time effort20%of usual time effortXMore than 20%of usual time effortMore than 20%of usual time effortXXNo time savingsNo time savings#IBMTechXchangeRetrieval Augmented Generati
43、on PatternUserQuestionSearch&RetrievalPrompt=Instructions+Search Results+QuestionLLMGenerated output with sourcesTop search resultsSee a real-life example of how IBM uses RAGto answer user questions about watsonx.ai inthe product documentation33IBM TechXchange|2024 IBM Corporation#IBMTechXchangeRetr
44、ieval Augmented Generation BenefitsSource of truthSource of truthRAG implementation can easily include the the sources of informationEasy for a human to verify the accuracyBusiness relevanceBusiness relevanceSpecify what documents are searched for answers to ensure that the LLM will respond using business-relevant informationInformation currencyInformation currencyMost current information is used to generate the responseAddress HallucinationAddress HallucinationMain weakness of LLMsOne of the best ways to minimize hallucinationsIBM TechXchange|2024 IBM Corporation34