利用 Mphasis 的 Agentic AI 框架推动现代化（由 Mphasis 赞助）.pdf

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Anup Nair He/Him CTO Mphasis.aiMphasisBharat KundnaniHe/Him Sr Architect/Partner Mphasis.aiMphasis

2、Driving Modernization Using Mphasis Agentic AI FrameworkM A M 2 1 9-S 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.“Every CIO tells me the same thing:I cant innovate because my core systems are too risky to touch.”We have 49 core systems.Touching any one of them breaks the othe

3、r 48.At Mphasis we decided to solve this problem 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.The Problem with Legacy SystemsLegacy systems anchor the enterpriseCritical intelligence lives in code,not in data knowledge is buriedDependency on high-end engineersFragmented operati

4、ons across business unitsEvery team re-invents the wheelChange is slow and costlyAI hits the wall.Without connected data or business context,AI models cannot scaleWe must move intelligence from code to connected,contextual dataOur ConclusionOur InsightIts not about outdated technology;its how busine

5、ss is encoded 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Legacy ModernizationRelearnExtract Business Rules and Understand 1RedefineDefine User Stories and Value streams2Re-architectDefine Target State Architecture3Re-codeWrite target state code,test,deploy4 2025,Amazon Web Se

6、rvices,Inc.or its affiliates.All rights reserved.Relearn AgentOntosphereBusiness Analyst AgentUser StoriesDomain ModelArchitecture AgentFlowsRules&DataArchitecture GuidanceEngineering AgentCodePackagingDeployment Topology,TelemetryRunOps AgentCodeDocumentsModernization Approach 2025,Amazon Web Servi