IBM Meetup 2026-3 – Know Your Agents, Trust Your Context

About this event

👋 Welcome back, IBM ATL Data & AI Meetup! This is going to be the last session for 2026, and we have a packed agenda.

We start with what we brought home from TechXchange in October, then dig into agents you can trust and the context they run on. Expect a sampler: whatever we do not reach moves to January, which doubles as a teaser for that session (date still being worked out).

📋 HIGH LEVEL FLOW Three parts: a welcome and recap, a lightning round, then a main thread where we go deeper on topics we have been building across meetups.

🏠 PART 1: WELCOME, RECAP AND WHAT IS NEW IN BOB 🛠️ TechXchange recap and Bob updates: Many know that IBM TechXchange, our technical conference, was in Atlanta at the end of October. As expected, there were several new announcements and updates. Today we focus on the Bob updates.

⚡ PART 2: LIGHTNING ROUND 🔐 Quantum-safe encryption, part two: A demo that picks up from last time's lightning talk. 🖥️ IBM Power, part one of two: Why Power at a Data and AI meetup? Enterprises want to own more of the stack under their AI, for privacy, control and token costs, and self-hosted Bob is one example. In January we look at how agentic AI tools like Bob shorten the learning curve. ⚖️ Deterministic and probabilistic decisions in agent orchestration: A concept talk, not a product pitch. Using Jev as one example of a small, fast decision model, we look at which decisions stay deterministic, which stay probabilistic, and how to combine them in an agent flow. 🎤 Your topic: Got five minutes and something to say? Tell us.

🧵 PART 3: MAIN THREAD, GOING DEEPER 🔑 Agent identity and control: When an agent acts, whose action is it? What changes when both the person and the agent have an identity that shows up in logs and traces, and what a kill switch for a misbehaving agent means for how much control you really have. 🔍 Agent discovery: Agents are being built on every major cloud and platform, and in code repositories. Which ones exist? Which are governed, observed, traced, logged, attributed or duplicated, and which are invisible to you? The ones nobody told you about are the real risk. 📚 From good retrieval to repeatable knowledge: Last time we showed retrieval working well. Can we do better? How to make knowledge, retrieval and prompts repeatable instead of hand-tuned, using a generative programming approach. 📡 Agent context in real time: A theme taking shape: agents that act on what is happening now, not last night's batch. A first look at streaming data as live context for agents, and what it changes about how we design data.

⏰ DRAFT RUNNING ORDER -> 1:00 PM Doors open | Networking and intros -> 1:10 PM Welcome and recap: TechXchange and what is new in Bob -> 1:25 PM Lightning round: 10 to 15 minutes, flexible -> 1:40 PM Main thread (AI): Agents you can trust, identity, control and discovery -> 2:00 PM Main thread (Data): Context for agents, repeatable knowledge and real-time streams -> 2:20 PM Open floor | Questions, debate, hallway conversations -> 2:30 PM Close This is a lot for 90 minutes, so expect a sampler. The main thread flexes to absorb timing, and whatever we do not reach moves to our January meetup. Final running order is posted closer to the day.

From IBM Data and AI Atlanta meetup