AI automation: Getting started with Red Hat Ansible Automation Platform

Imagine you’re running a busy IT team. Every day, there are repetitive chores—installing software, scaling out solutions, checking system health or pushing out security patches. These challenges and more are part of the IT practitioner’s day-to-day life. Now your CTO comes to you and tells you that she needs the new AI infrastructure and applications moved from experimental proof of concept into production. You have no new headcount or resources to help support these efforts, and the end solution needs to integrate with existing business-critical workflows.

Automation to the rescue

When standardization is part of your enterprise IT strategy, you can use automation for AI infrastructure to help address this added workload. Some of you are already using Red Hat Ansible Automation Platform, so you'll be able to tackle AI challenges using skills you’ve already developed.

For those just getting started with automation, however, don’t worry. You can get started quickly using Ansible Automation Platform and even use the generative AI (gen AI) feature Red Hat Ansible Lightspeed to help reduce the learning curve.

Regardless of where you are with automation skills, applying automation to IT (and AI is really just more IT) will help reduce costs while accelerating your time to achieving AI value. 

So how do you get started achieving AI value? The first step I recommend probably won't surprise you: automate your AI deployments! Red Hat refers to this as "AI infrastructure automation," and it provides measurable business value from overall time savings, reduced errors and improved system reliability.

How to automate your AI infrastructure

Using Ansible Automation Platform to automate the installation, configuration and maintenance of Red Hat OpenShift AI and Red Hat Enterprise Linux AI (RHEL AI), you can deploy both predictive and gen AI solutions, reduce manual tasks and provide consistent configurations and optimizations across AI deployments.

Although AI infrastructure delivers huge end-user value, the IT operations requirements are still largely manual. Some examples of automations that will help you get started deploying AI solutions on RHEL AI and OpenShift AI include:

• Establishing secure connectivity between systems—including edge—used for transferring and accessing the data that drive AI solutions
• Standardizing deployments to promote uniformity and reliability of both RHEL AI and OpenShift AI using Ansible Playbooks
• Automating security so users only have access to the data and systems they are authorized to access

Additionally, other infrastructure for AI solutions benefit from automation:

• Hardware, such as top-of-rack network switches and routers, still needs to be installed and configured
• Vector databases for use with retrieval augmented generation (RAG) solutions need to be installed and managed
• Load balancing hardware and software are needed for HTTP access to AI models using OpenAI, or APIs like Llama Stack
• Connectivity and storage for model training and alignment data needs to be set up and managed

And there are many more use cases that still need to be automated. That’s a lot to handle, but many deployment topologies will include components that can be automated, like the following:

Leverage Event-Driven Ansible to build out AIOps workflows

OK, automation of AI infrastructure… check! So what’s next? As mentioned earlier, the automation skills you may already have can be applied to AI technologies that now live on or connect to that AI infrastructure. However, what about new skill requirements, or simply the desire to increase automation productivity overall? Can AI help there? Of course!

Using Red Hat Ansible Lightspeed with IBM watsonx Code Assistant, automation developers get the combination of gen AI capabilities and robust automation tools that allow you and your team to work smarter and deploy solutions faster, simplifying the journey toward comprehensive AI infrastructure automation. Plus, if you’re using OpenShift AI, you’re already using Red Hat OpenShift, which can run many other workloads (including VM virtualization and management) which Ansible Lightspeed can also help you automate.

It’s important to note that gen AI developer tools built into Ansible Lightspeed aren’t just a simple “check the box” AI solution. Red Hat has thoughtfully integrated generative and predictive AI into automation workflows seamlessly, so users may not even know that they're using it. This allows Ansible Automation Platform content creators to develop automation code faster, more efficiently and with higher accuracy.

Fully automating AIOps workflows

So far, everything in this article has involved a human in the loop. But can automation magically take action without human intervention? Absolutely!

A proper AI operations (AIOps) workflow is powered by event-driven automation, which automatically performs operations triggered by events that occur in your existing IT deployments. When put in place, event-driven automation helps AIOps deliver the following benefits:

• Resolution speed: AIOps can help reduce downtime by detecting and reacting to emerging issues, decreasing mean time to resolution (MTTR).
• Self-healing or closed-loop automation systems: It enables self-healing infrastructure, which can significantly improve performance and uptime.
• Big data: AIOps can put big data to use by more efficiently cleaning, analyzing and taking action on it.
• Efficiency and scale: It can increase staff efficiency by using insights from AI models to identify actions and scale detection.
• Simplification: AIOps can streamline and potentially fully automate many repetitive IT service management tasks.
• Real-time data correlation and decision making: When AIOps includes an automation engine, it can respond automatically based on data—reducing human intervention and error while minimizing noise.
• Scaled data correlation and prediction: AIOps can automatically analyze significantly more data,  far beyond what humans can do manually.

Adding business and security policies to safeguard AI workflows

Finally, after you deploy and automate your AI infrastructure, after you successfully integrate the AIOps workflows and you now have full self-healing closed-loop automation, you now need guardrails to make sure the AI only functions as it’s intended.

Your AI strategy needs to be trustworthy, reliable and accountable to the business policies you’ve had for years. Policy enforcement is the last step in this story—what you need are guardrails for your AI decisions to make sure that even though something could be automated, it doesn't mean it should be automated without the required compliance and security checks.

Learn more

You can learn more about the benefits of choosing Red Hat Ansible Automation Platform for your AI foundation in this article. If you want to dig into AIOps a bit more, take a look at this short video and this blog.

There are also two Red Hat Interactive experiences you can review in less than 2 minutes each: Red Hat Ansible unlocks AIOps and Event-Driven Ansible with Red Hat OpenShift.