2026 GET-AI SERIES: 6 — Enabling Autonomous Operations
🅿️ Free Parking: Register at the lobby kiosk and enter Gora Datta or Brian Hagerty when prompted. They will be notified and escort you to the meeting room. 🍽️ Dinner will be provided for in-person attendees.
Artificial Intelligence is rapidly progressing from systems that generate responses to intelligent agents capable of reasoning, planning, using tools, maintaining context, and executing multi-step tasks. As organizations move toward agentic AI, the opportunity is no longer limited to helping employees find information—it is increasingly about enabling AI systems to participate in and automate real business processes.
Following our August session on Connecting AI to Enterprise Data using Model Context Protocol (MCP), the IEEE Orange County Computer Society GET-AI Series continues in September with a practical, demo-driven exploration of AI Agents and Autonomous Operations.
This double-feature session begins with the foundations of agentic AI—what makes an AI system an agent, how agents reason and plan, how they use tools, memory, and context, and how to build a simple agent capable of completing a multi-step task. The second session builds on those concepts to explore how autonomous and multi-agent systems can orchestrate complex enterprise workflows while maintaining security, governance, observability, human oversight, and operational reliability.
By combining foundational concepts with architecture patterns and live demonstrations, the September session will help attendees understand how organizations can progress from AI assistance to intelligent, governed autonomous operations.
Session 1: AI Agents 101: Build Your First Intelligent Agent
45 minutes
Through practical explanations and a live demonstration, this session introduces the fundamentals of Agentic AI and explains how AI agents differ from traditional chatbots and standalone LLM applications.
Attendees will explore the core building blocks of an AI agent, including reasoning, planning, memory, context, tools, and actions. The session will then demonstrate how these components work together by building a simple agent capable of interpreting a goal, selecting appropriate tools, executing multiple steps, and returning a meaningful result.
Topics Include
- What Is an AI Agent? From Assistants to Agentic AI
- Core Agent Components: Reasoning, Planning, Memory & Actions
- Tools, Function Calling & Context Management
- Human-in-the-Loop Controls and Common Agent Design Patterns
- Build and Demonstrate a Simple Multi-Step AI Agent
👉 Takeaway
Understand how AI agents reason, plan, use tools, and take actions—and learn how to build a basic intelligent agent capable of executing a multi-step task.
Session 2: Autonomous Enterprise AI: Orchestrating Agents, Tools & Business Workflows
Building on the agent foundations introduced in Session 1, this session explores what it takes to move from a single AI agent to reliable, governed autonomous enterprise operations.
Attendees will examine architecture patterns for orchestrating agents across applications, APIs, enterprise data, and business processes. The session will explore when to use a single agent versus multiple specialized agents, how agents collaborate and delegate tasks, and how organizations can maintain control through approvals, identity, policy, observability, evaluation, and operational guardrails.
The session will also examine the production challenges that arise when agents begin taking actions—including incorrect tool selection, uncontrolled loops, unreliable outputs, excessive autonomy, and failure recovery—and practical strategies for addressing them.
A live demonstration can showcase an end-to-end autonomous workflow in which an agent receives a business objective, gathers context, selects tools, performs multiple steps, requests approval when appropriate, and completes the workflow.
Topics Include
- From AI Agents to Autonomous Operations
- Enterprise Agent Architecture: Single-Agent & Multi-Agent Systems
- Agent Orchestration, Delegation & Enterprise Tool Integration
- Governance, Human Oversight, Observability & Failure Recovery
- Production-Ready Agentic Systems with Live End-to-End Demonstration
👉 Takeaway
Learn how to design secure, observable, and governed agentic systems that coordinate tools, data, and business processes to enable reliable autonomous enterprise operations.
About the Organizer
Pradyumna Kodgi
Principal Product Manager | Oracle Health & AI
IEEE Senior Member | Vice Chair, IEEE EMBS – Orange County
Member, IEEE AI Agentic Systems & AI Policy Committees
📍 California, USA
📧 pkodgi@ieee.org
🔗 linkedin.com/in/pkodgi
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
Loading virtual attendance info...
- 5270 California Ave
- 2nd Floor
- IRVINE, California
- United States 92617
- Building: Beall Applied Innovation
- Room Number: Emerald Cove Conference Room
- Contact Event Hosts
- Co-sponsored by Pradyumna Kodgi
- Starts 01 September 2026 07:00 AM UTC
- Ends 28 September 2026 07:00 PM UTC
- No Admission Charge
Speakers
Brij
Session 1: AI Agents 101: Build Your First Intelligent Agent
Everyone is talking about AI agents, but far fewer people can clearly explain what makes an agent different from a chatbot with extra steps.
This session answers that question by building an agent live. It begins with a practical definition: an agent is a model combined with tools, a reasoning loop, and a stopping rule. From there, the session explores the core components that make agents work, including reasoning, planning, memory, context, actions, and function calling.
The live demonstration uses a synthetic retirement-planning scenario. Given a sample client profile, the agent identifies a projected shortfall and independently explores possible strategies such as retiring later, increasing contributions, reducing expenses, or combining multiple approaches. Along the way, the agent validates assumptions against explicit guardrails, avoids recommending specific financial products, and pauses for human review before any result is finalized.
The session will also examine how to test agentic systems that may not behave identically on every run.
The demonstration is intentionally framework-free, using only the Python standard library and one API client, so attendees can clearly see the underlying agent mechanism rather than an abstraction provided by an orchestration framework. All data is synthetic and all figures are illustrative. Nothing presented constitutes financial advice. Attendees are encouraged to bring a laptop and follow along; the code and notebook will be made publicly available.
Biography:
Brij Mohan is an independent researcher specializing in agentic AI, multi-agent system design, and cloud-native data architecture. He brings more than 25 years of engineering and leadership experience across fintech, healthcare, and government systems, including previous experience as a founder and CEO.
His current research focuses on multi-agent architectures for fraud detection, master data management, and entity resolution. Brij is an IEEE Senior Member, ACM Professional Member, a member of the Forbes Technology Council, and serves on the Products That Count Advisory Council.
He also holds the AWS Solutions Architect Professional and AWS Machine Learning Engineer Associate certifications, along with PMP, PMI-ACP, and TOGAF 9 credentials.
Email:
Address:United States
Karthik of IBM
Session 2: Autonomous Enterprise AI: Orchestrating Agents, Tools & Business Workflows
Most enterprise AI today is assistive: a person asks, the model responds, and the person decides. Autonomous operations changes that model by allowing agents to handle the normal path while humans focus on exceptions.
This session explores what it takes to make that shift in real enterprise environments. It examines how agents interact with growing catalogs of tools and systems, how orchestration can fail silently when the wrong capability is selected, and why identity, authority, auditability, and independent verification are essential for production use.
The session will also cover how organizations can design reliable autonomous workflows, recover from partial failures, maintain human oversight where needed, and measure success based on how much additional demand a business function can absorb without increasing human processing effort.
Biography:
Karthik Karunanithi is a Solution Delivery Leader for AI Transformation at IBM, where he designs production-grade agentic AI systems for enterprise and public-sector clients. His work includes multi-agent architectures, Model Context Protocol integrations, and the identity and reliability layers required to operate these systems in regulated environments.
He is a named inventor on two granted U.S. patents in secure access control for patient health information, a Senior Member of IEEE, and a voting member of the IEEE P7999 working group on AI ethics oversight.
Karthik also writes about production agent systems at karthikk.hashnode.dev.
Email:
Address:United States
Agenda
September Session Agenda
Following the same format as August:
4:30 PM – 5:00 PM — Registration, Check-In & Networking
5:00 PM – 5:15 PM — Welcome, Introductions & Opening Remarks
Google Meet opens at 5:00 PM. Virtual attendees may join at this time.
5:15 PM – 6:00 PM — Session 1: AI Agents 101 — Build Your First Intelligent Agent
6:00 PM – 6:15 PM — Dinner & Networking
6:15 PM – 7:00 PM — Session 2: Autonomous Enterprise AI — Orchestrating Agents, Tools & Business Workflows