Ashutosh Rana is an Enterprise Architecture and Software Development leader with more than 20 years of experience delivering large-scale digital transformation across CRM, ERP, HRMS, and student-lifecycle systems. He specializes in modernizing enterprise applications with Salesforce, PeopleSoft, and Google Cloud Platform (GCP), and in integrating complex ecosystems through MuleSoft, IBM MQ, REST APIs, Integration Broker, and AI-driven technologies. Across the education, telecom, and finance sectors, Ashutosh has architected AI-powered assistive and autonomous agents using large language models (LLMs) and Retrieval-Augmented Generation (RAG), built unified customer and student data platforms on BigQuery and Vertex AI, and delivered omni-channel communication systems spanning chat, SMS, WhatsApp, and voice. He has also led multiple PeopleSoft-to-Salesforce CRM migrations for universities and automation initiatives that measurably improved operational efficiency and user engagement. Currently Lead Developer and Technical Architect at CherryRoad Technologies / Infosemantics Inc., he drives CRM modernization, DevOps automation, identity federation, and assistive and autonomous agent solutions built on cloud-based integration. He previously held architecture and consulting roles at IBM, Hexaware Technologies, and Oracle India, contributing to PeopleSoft CRM product development, enterprise integrations, reporting frameworks, and large-scale business transformation programs. Ashutosh holds a Master of Computer Applications (MCA) from Guru Gobind Singh Indraprastha University, Delhi, and multiple Salesforce certifications — including Salesforce Application Architect and Integration Architecture Designer — along with Oracle certifications in consulting and advanced development. An IEEE Senior Member, he is an active contributor to the technology community through publications, conference speaking engagements, and professional memberships.

Title:Governed Multi-Agent AI Systems: Compliance, Auditability, and Human Oversight at Scale -  Abstract:As enterprises move from single-purpose chatbots to orchestrated fleets of assistive and autonomous AI agents, the hardest challenges are no longer about model capability — they are about governance. When multiple agents plan, call tools, exchange messages, and act on real customer and student data, organizations must answer difficult questions: Who authorized this action? What data did the agent see? Can every decision be reproduced and audited? And when must a human remain in the loop? This session presents a practical architecture for governed multi-agent AI systems that can be trusted in regulated sectors such as education, telecom, and finance. Drawing on large-scale implementations across CRM, ERP, and student-lifecycle platforms, the talk outlines a reference design that layers policy enforcement, identity and access controls, and Retrieval-Augmented Generation (RAG) grounding onto agent orchestration. Attendees will learn how to build end-to-end auditability — capturing prompts, tool calls, data lineage, and outcomes — while embedding human oversight through approval gates, guardrails, and escalation paths. The session translates emerging governance principles into concrete engineering patterns, leaving participants with an actionable blueprint for deploying multi-agent AI at enterprise scale while maintaining compliance, transparency, and accountability.