GenAI 從數據開始：創建基于強大數據核心的聊天機器人.pdf

《GenAI 從數據開始：創建基于強大數據核心的聊天機器人.pdf》由會員分享，可在線閱讀，更多相關《GenAI 從數據開始：創建基于強大數據核心的聊天機器人.pdf（25頁珍藏版）》請在三個皮匠報告上搜索。

1、GenAI Starts with DataCreate a Chatbot Built on a Strong Data CoreSPEAKERSKevin BohanDirector of Product MarketingRomeel ShethData EngineerAGENDA1.What kind of solution is Denodo?2.How does Denodo leverage AI?3.Show me!4Denodo Platform:One-stop-shop for a data pipelineSelf-ServiceHybrid/Multi-CloudD

2、enodoAssistantQueryOptimizationData Security andGovernanceEmbedded MPP EngineINTEGRATEdisparate data in any location,format or latencyMANAGErelated data into views with universal semantic modelDELIVERusing BI&data science tools,data catalog,and APIsLOGICAL DATA MANAGEMENTFinOpsDATA AS A SERVICERESTf

3、ul/ODataGraphQL/GeoJSONBI ToolsData ScienceToolsDATA CATALOGDiscover ExploreDocumentCONSUMERSSQLAI AgentsAI SDKSOURCES200+dataadaptersAppsStreamingSaaSFilesLakehousesOpen Table formatsTraditionalDB&DW5What is a Logical Data Farbic?6Why Denodo AI RAG?Isnt every solution using some form of RAG?Of cour

4、se-everyone is using the same LLM technology to generate SQL So why should I care about Query RAG?A successful RAG implementation isnt about which LLM you choose or how fast you can load a vector store Its about METADATA!At Denodo,metadata management is our core competency!7Denodo AssistantBy levera

5、ging internal algorithms developed by Denodo and integrating with LLMs offered by industry leaders,Denodo Assistant helps developers and users throughout the data product life cycle by intelligently answering questions,offering suggestions,and recommending optimal solutions.SUGGESTASKRECOMMEND8Can D

6、enodo Assistant Enrich Metadata?9Can Denodo Assistant Debug My Queries?ExplainBusiness purpose of a SQL queryDiagnose Automatically detect and fix typos,syntax and grammatical errors 10Can Denodo Assistant Help Build My Model?11 Natural language SQL queries Recommends relevant datasets Simplifies da

7、ta customization with usage-based suggestions for filters and data transformations Auto-generates view and column descriptions Recommend business-friendly column names Optimizes query performance with aggregate aware summary suggestions SQL statement explanations Debugging of SQL syntax and grammar

8、errors Context-aware VQL autocompletion LLM-based text enrichment in SQLDenodo Assistants Main Features12How does GenAI Adoption Fail?How many loans have been granted this week?Sorry,I dont have access to specific information about loansHow many loans have been granted this week?234 loans were grant

9、ed this weekCorporate DataAugment LLM with Corporate Data and Knowledge13Why Exactly is GenAI Adoption Complex?Whats my current balance?Where is the package I ordered last week?Whats the current stock of item XYZ?Has supplier Y paid their latest billHow does this quarters forecast compare with last

10、year?Payments applicationUPS,Fedex APIsERPPayments applicationCRMEDWInformation for these kind of genAI applications is often scattered in a variety of operational and analytical systems,both external and internal14How many loans have been granted this week?Vector DBIntroducing Denodo AI SDKSELECT c

11、ount(*)FROM loansWHERE get_week(grant_time)=get_week(now()Denodo AI SDK provides:Encapsulate the process in an easy-to-use SDK Simplifies genAI adoption for developers234 loans were granted this weekDenodo AI SDK15Denodo AI SDK ArchitectureSOURCESLOGICAL DATA FABRICAI AGENTLLM OrchestrationVector Da

12、tabaseLarge Language ModelDenodo AssistantData Security andGovernanceQueryOptimizationSemantic LayerAI SDKAmazon OpenSearch200+dataadaptersAppsStreamingSaaSFilesLakehousesOpen Table formatsTraditionalDB&DW16How Can You Use Denodos Chatbot?17Build Third-Party Integrations in Minutes18DemonstrationGen

13、AI Starts with Data:Create a Chatbot Built on a Strong Data Core19Business Use CaseScenario We are an e-commerce company.We track both customer and sales data.We have disparate data sources that need to be leveraged together.We want to leverage our existing tools(database and visualization).Executiv

14、es need high-level visibility over our top spenders.We have users that are familiar with and new to SQL and database technology.We have security policies over sensitive,personal data.Report performance needs to be quick,as stakeholders will review them on a weekly basis.End users need to explore our

15、 data.We need a more agile architecture and empowers our metadata.Goal Configure a logical data fabric through Denodo.AGENDA1.Connect and Combine Data Sources2.Performance and Optimization3.Securing your Data Fabric4.Publishing5.Self-Service with the Data Catalog6.Chatbot7.Model Consumption in Power

16、BI21Speaker Kevin Bohan Director of Product Marketing,DenodoTitle:From Data Chaos to AI Clarity:Strategies to Supercharge Your AI InitiativesAbstract:Data is the fuel for Generative AIbut what happens when that fuel is scattered across disconnected systems,buried in legacy applications,or locked beh

17、ind security constraints?Many organizations struggle to harness the full potential of their data,limiting the effectiveness of AI initiatives.In this session,well explore how leading enterprises have broken down these barriers to unlock real-time,context-aware insights for their AI-driven applicatio

18、ns.Youll learn how theyve tackled key challenges such as:Providing seamless access to distributed data without duplication or delays Ensuring governance and security while enabling AI-driven innovation Delivering high-quality,fact-based responses to enhance AI outputsJoin us to discover practical st

19、rategies that can help you unleash the power of your dataso your AI initiatives deliver real business impact,not just theoretical promise.Where else can you find us?22Where else can you find us?23Follow these steps to get started!1.Download Denodo Express if you dont have Enterprise Plus yet2.Read t

20、he step-by-step tutorial Building an AI Chatbot with Denodo in Minutes3.Review the documentation Denodo AI SDK Guide4.Dive into the open-source code of Denodo AI SDKin Github Curious about more advanced uses of the technology?Check out these KB articles:Adding session memory to your Denodo AI SDK ch

21、atbotIntegrating the Denodo AI SDK in a Slack botIntegrating the Denodo AI SDK in a MS Teams botIntrigued about the possibilities of Denodos AI SDK?Try it yourself!24Next StepsLearn more about Denodo Platform!Download Denodo Express for free:http:/ the Denodo Basics,Data Combination,Data Services,Da

22、ta Consumption and Using Data Catalog tutorials.Take Denodo for a Test Drive:Take the Agile BI and Analytics test drive.Read the KB articles:Denodo Cloud Architectures and Denodo Best Practices and GuidelinesThanks! Copyright Denodo Technologies.All rights reservedUnless otherwise specified,no part of this PDF file may be reproduced or utilized in any for or by any means,electronic or mechanical,including photocopying and microfilm,without prior the written authorization from Denodo Technologies.