Red Hat: Building Omnichannel AI Agents with OpenShift AI

Name: Red Hat: Building Omnichannel AI Agents with OpenShift AI
Uploaded: 2026-07-01T15:45:36-04:00
Duration: 3 min 26 s
Description: Looking to automate complex IT service tasks? Learn how to build an omnichannel self-service agent using Red Hat OpenShift AI that meets your team where they already work: Slack, email, and ServiceNow. In this video, Tommy Hughes, Principal Software En...

Red Hat

07/01/2026

0 (0%)

Report Like Favorite

Transcript

workflow that you can run out of the box. We'll show you how a conversation can start in Slack and continue in email. Same session, across channels, without losing context. Then we'll show you the architecture behind it. With this QuickStart, you get a template that you can extend with your own channels, MCP servers, and agents. Let's get started by asking the agent for some help via our Slack channel. You see the agent has responded. Our routing agent is asking us what we need. I tell it I want to refresh my laptop. The platform is determining my email address from my Slack profile, and we'll use this throughout the rest of the process. It's returned my details, including the current laptop model I have, and it tells me that I'm eligible for an upgrade. Asks me if I'd like to proceed, I say yes. Now let's switch over to email, where we see the same conversation with our agent is occurring. Looks like a new response has come through. We'll review it, and see that it's asking us to select a new laptop. Let's go with option three, the ThinkPad. We can reply from here with our selection. Let's go ahead and finish our conversation in Slack. Now the agent asks if we'd like to proceed with creating a ticket in ServiceNow. We'll say yes. Great. Now let's verify that the ticket actually exists in our ServiceNow instance. We'll refresh, and here's that request in ServiceNow. Let's verify that the details match, and they do. Now let's look at the architecture. Users reach the system through whatever channels are configured. Currently this can include Slack, email, or the API, including the CLI. The request manager receives from those channels, validates and routes every request. It ties every request to the same user session, so switching channels keeps one conversation. And it then delivers responses back to the configured channels. This is deployed to OpenShift using elements of OpenShift AI. The components all talk with cloud events over Knative eventing with Kafka, which makes it event-driven and scalable. The routing agent figures out what the user needs and hands off to a specialist agent. In our demo, that was a laptop refresh agent. That agent uses knowledge bases and tools like ServiceNow via MCP. The QuickStart is built to be easily extensible. You can add more channels, like Discord, Teams, or SMS. More IT services, like access management or compliance workflows. Or new specialist agents, all without changing the core platform. With this QuickStart, you get one conversation across all channels, and it's ready to extend. Thanks. ♪AT&T jingle♪

TL;DR

Red Hat's IT Self-Service Agent QuickStart enables omnichannel AI agents that maintain conversation context across Slack, email, and API interfaces, allowing users to switch channels mid-conversation without losing session state.
The architecture uses OpenShift AI with event-driven communication via Knative eventing and Kafka, featuring a request manager that routes to specialized agents and integrates with enterprise systems like ServiceNow through MCP.
The QuickStart template is built for extensibility, allowing organizations to add new channels, IT services, or specialist agents without changing the core platform, providing a production-ready foundation for enterprise AI automation.

Summary

This demonstration showcases Red Hat's IT Self-Service Agent AI QuickStart, a ready-to-deploy solution that enables organizations to build intelligent, omnichannel service agents using OpenShift AI. The demo walks through a practical laptop refresh workflow where users can seamlessly interact with an AI agent across multiple channels—starting a request in Slack and continuing it via email—while maintaining full conversation context throughout the session. The architecture leverages OpenShift AI components with event-driven communication through Knative eventing and Kafka, enabling scalable and decoupled service management. The QuickStart template is designed for extensibility, allowing teams to add new communication channels, integrate additional IT services, or deploy specialized agents without modifying the core platform. This approach demonstrates how enterprises can deploy production-ready AI agents on open source infrastructure while maintaining the flexibility to customize workflows for their specific IT service management needs.

Chapters

0:00 - Introduction and Demo Overview
0:28 - Laptop Refresh Workflow in Slack
1:04 - Channel Switch to Email
1:41 - ServiceNow Integration Verification
2:01 - Architecture Deep Dive

Key Quotes

0:16 "Same session, across channels, without losing context."
2:19 "It ties every request to the same user session, so switching channels keeps one conversation."
2:53 "The QuickStart is built to be easily extensible."

FAQ

How does the agent maintain conversation context when users switch between channels like Slack and email?

The request manager component ties every request to the same user session regardless of channel, validating and routing requests while maintaining conversation state. This allows users to start a conversation in Slack and continue it via email without losing context or having to repeat information.

What makes the IT Self-Service Agent QuickStart extensible for different use cases?

The QuickStart is designed with a modular architecture that allows teams to add new communication channels (Discord, Teams, SMS), integrate additional IT services (access management, compliance workflows), or deploy new specialist agents—all without modifying the core platform components.

Categories:

Tags:

Show more Show less

TL;DR

Summary

Chapters

Key Quotes

FAQ

Becoming Agent Ready with Cyera: Essential Strategies and Insights

Verge.io: Can You Afford Your Next Storage Refresh?

Summer of Satori: FunFoneFarm's Transformation of Fraud into Seamless Integration