PILLAR 1 · PLATFORM · THE AGENTIC AUTOMATION FRAMEWORK BY LSD OPEN
Turn AI that chats into AI that acts
Private, sovereign agentic AI built on Red Hat.
LSD Open’s Agentic Automation Framework combines Red Hat OpenShift AI, Ansible Automation Platform and a large language model such as Llama Stack, AWS Bedrock or other models, so agents can plan, reason and act on your infrastructure.

Red Hat
OpenShift AI and Ansible Automation Platform
MCP
Open standard connecting agents to your tools
Private
Runs in your own datacentre
Zero-touch
Incidents detected, diagnosed and fixed
FROM PASSIVE TO ACTIVE AI
From chat to action
AI shouldn’t be a standalone tool; it should be a core component of your platform engineering strategy. The agent plans and reasons, MCP connects it to your systems, and Ansible Automation Platform does the work.
01
AI agent plans and reasons
Payments pods keep crashing. Fix it.
Plan: check logs, find the cause, run the right fix, verify.
Red Hat OpenShift AI + vLLM, Llama Stack / AWS Bedrock / other models
02
Model Context Protocol (MCP)
APIsAutomationData sourcesOne standard interface to your existing systems
Standardises the interface to existing APIs, automation systems and data sources
03
Ansible Automation Platform
✓Check pod logs
✓Raise memory limit
✓Rolling restart
✓Verify health
Deterministic automation tasks exposed as tools
04
Your infrastructure
Incident fixed. No human in the loop.
Incidents detected, diagnosed and fixed
WATCH THE OVERVIEW
See the framework in action
Watch the video below for a brief overview:
AN ILLUSTRATIVE EXAMPLE
How an incident gets fixed.
Existing Ansible playbooks become tools the agent can call. The agent decides which one to run, Ansible Automation Platform runs it, and the result is checked before the incident is closed.
Payments API latency
Latency highPods restartingYour OpenShift platform
pod
pod
pod
pod
pod
pod
Monitoring raises an incident on your platform.
Evidence gathered through MCP
payments-api-1: started
payments-api-1: OOMKilled (limit 256Mi)
payments-api-4: started
payments-api-4: OOMKilled (limit 256Mi)
metrics: memory climbing past limit
Agent reasoning
Reading logs and metricsCause foundMemory limit too low for current load. Needs a config change, then a safe restart.
Ansible playbooks exposed as tools
Restart podplaybook
Rotate certificatesplaybook
Raise memory limit and roll restartplaybook
Scale deploymentplaybook
Agent decision
Matching cause to toolExisting automation you already trust becomes a tool the agent can call. It chooses one.
Ansible Automation Platform job
✓Check pod logs
✓Raise memory limit
✓Rolling restart
✓Verify health
Your OpenShift platform
pod
pod
pod
pod
pod
pod
Deterministic, repeatable and logged.
Payments API latency
Latency normalPods healthyYour OpenShift platform
pod
pod
pod
pod
pod
pod
Incident closed. The agent checks the outcome, records what it did and hands the audit trail to your team.
Why Agentic Automation?
While the enterprise world is scrambling to find suitable use cases for AI in their line of business, many AI implementations are unsuccessful because they aren’t delivering the value to the business. With agentic AI, automation, and scalable cloud native platforms, this framework creates powerful tooling for your team to use in day-to-day operations, freeing their time from toil, admin, and repetitive tasks.
Sovereignty & Governance
Deploy AI workloads directly within your own datacenter. Keep your data private and align strictly with your existing governance and protection requirements. No data leaks. No public cloud dependencies.
Action Over Chat
Stop building chatbots and start building agents. LSD Open shifts the paradigm to Agentic AI – intelligent models capable of planning, reasoning, and executing complex tasks autonomously.
Powerful Tooling for AI
Existing Ansible automations now become tools for AI agents to use, enabling teams to rapidly build out custom AI tools using Ansible playbooks, and integrate them using model context protocol.
Zero-Touch Operations
Reduce Mean-Time-To-Resolution (MTTR) by allowing agents to detect, diagnose, and fix incidents without human intervention.
Economic Control
Shift from volatile SaaS OpEx models to predictable infrastructure costs.
Foundation for Enterprise AI
Build on a platform that scales from first agent to enterprise-wide AI, on the Red Hat stack your teams already run.
BUILT ON RED HAT
The Agentic Automation Framework
The Agentic Automation framework is a private, sovereign AI capability hosted entirely within your datacenter. By combining the reasoning power of Red Hat OpenShift AI with the deterministic reliability of Ansible Automation Platform, the framework enables a digital workforce of digital agents capable of autonomous operations. By using specific automations, model context protocol (MCP) and Large Language Models (LLMs), AI can add actual value to your business.
Agent and inference
Agent framework
Llama Stack / AWS Bedrock / other models
Inference
Red Hat OpenShift AI + vLLM
Integration and automation
Integration fabric
Model Context Protocol (MCP)
Automation
Ansible Automation Platform
Foundation
Foundation
Red Hat OpenShift
ScalabilityGPU scheduling
Framework Core Components
Red Hat OpenShift
Cloud native foundation for all components, enabling scalability and GPU scheduling.
Red Hat OpenShift AI + vLLM
Model serving and orchestration for open source LLMs, running on your own OpenShift.
Ansible Automation Platform
Executes deterministic automation tasks exposed to the agent as tools via MCP.
Model Context Protocol (MCP)
Standardises the interface between LLM agents and existing APIs, automation systems and data sources.
Agent framework: Llama Stack / AWS Bedrock / other models
LLM capable of agent orchestration, reasoning and tool integration.
Your existing automation
Existing Ansible playbooks become tools the agent can use, so the work you already trust is not rebuilt.