Building Context-Aware AI Agents with Real-Time Data

Опис

A free online ODSC AI Dublin webinar on 9 September 2026 shows how IBM, Confluent and Google Cloud build context-aware AI agents with live data.Hosted by [ODSC AI Dublin](https://www.meetup.com/dublin-data-science-odsc/), this AI Skills Accelerator session begins at **6:00 PM** and focuses on the architectural layer often missing from agentic AI systems: continuous, refined context. Rather than asking an agent to act on stale or unprocessed information, the speakers will demonstrate how streaming operational data can support timely, reliable decisions.## What the session coversExperts from IBM, Confluent and Google Cloud will explain how their technologies work together to turn changing business data into useful context for AI agents. The session will explore:- How Confluent’s data streaming platform transforms operational data into automated decisions- Why fresh context matters when moving from brittle pipelines and prototypes to production-grade AI- How Google Cloud can add real-time investigation and explanation to streaming decisions- Connecting agents and data systems through MCP- Syncing governed data to BigQuery for historical records and analysis- Streaming-native patterns intended to improve reliability while reducing total cost of ownershipThe centrepiece is a **live anomaly-detection and investigation demo** built around racing telemetry. Data will stream through Confluent’s Real-Time Context Engine, where a Confluent Streaming Agent detects anomalies and recommends pit timing as conditions change. A [Gemini](https://www.createwith.com/tool/gemini) agent running in Google Antigravity will then investigate those decisions and explain the reasoning behind them.## Speakers- **Karan Sachdeva**, Global Lead for the IBM–Google strategic partnership at IBM- **Sean Falconer**, Head of AI at Confluent- **Sohrab Rahimi**, Director and AI/ML Lead at Google Cloud…[Register for this event](https://www.meetup.com/dublin-data-science-odsc/events/316073110/?eventOrigin=group_upcoming_events)