Making Complex Things Simple: Nutanix, Kubernetes, and Enterprise AI — Day 1, Part 1

By John Hein

Make the complex things simple. That theme captures what I find most valuable about the evolution of enterprise infrastructure.

Traditional virtual machines remain essential. Containers introduce different deployment and operating practices. AI brings models, inference services, and new questions about how applications use business information.

Thank you to Sonali Mishra, Principal Product Manager – NKP at Nutanix, one of our presenters, for sharing her perspective and expertise.

From my perspective as a service provider, the opportunity is bringing these requirements together in a way that customers and operations teams can actually manage. The product capabilities discussed below are drawn from publicly available Nutanix announcements and documentation, alongside my own observations.

One Operating Approach for VMs and Containers

Managing VMs and containers through a more consistent platform is an appealing proposition.

Customers have existing applications to support while exploring Kubernetes and AI. Their operational responsibilities grow with each new environment, making consistency increasingly valuable.

Nutanix Kubernetes Platform (NKP) 2.19 expands the options for managing virtual machines and deploying Kubernetes on physical servers through NKP Metal.

The practical benefit I see is flexibility: choose infrastructure that fits the workload while working toward common operational practices. Customers can continue supporting established applications while creating room for new ones.

Private Chatbots Need Business Context

Private chatbots offer a useful example of how AI can deliver practical value. Imagine an assistant that helps employees find answers in their own product documentation, support articles, and business information.

A model needs relevant context to answer those questions. That is where retrieval-augmented generation (RAG) comes in.

A typical RAG application retrieves relevant information from selected sources, adds it to the user’s question, and sends that combined context to an inference service. The model then generates a response informed by that material.

This approach does not require retraining the model every time a document changes. The application can retrieve updated information when answering a question, provided its underlying content and indexes are kept current.

For example, a product support assistant could retrieve the relevant section of a technical guide and use it to help explain a feature or troubleshooting procedure. The value comes from connecting the model to information that matters to that business.

Where Nutanix Enterprise AI Fits

Nutanix Enterprise AI (NAI) provides inference endpoints that applications can consume, including OpenAI-compatible APIs.

“OpenAI-compatible” describes the interface. It does not, by itself, mean the application sends its data to OpenAI.

That distinction is useful when discussing private AI applications. Developers can work with a familiar API approach while connecting their applications to the endpoints selected for their environment.

For a service provider, this opens an interesting discussion about delivering AI services alongside the infrastructure and data customers already depend on. The conversation becomes much more concrete when it starts with a business need, such as helping employees find accurate product information faster.

Making the AI Toolset Easier to Use

An AI application needs more than a model. Teams also need tools for development, experimentation, and data processing.

NKP 2.19 includes a curated AI application catalog for Pro and Ultimate environments, with tools such as JupyterHub and Kubeflow Pipelines.

The value of a catalog is reducing the integration work engineers must do before they can begin building something useful. A more consistent starting point gives teams more time to focus on the application and the people who will use it.

Making Product Knowledge Easier to Find

AI can also help the people learning and using the platform.

NKP AI Navigator provides answers using NKP documentation, Nutanix support knowledge, and upstream Kubernetes documentation.

For someone working through an unfamiliar feature, being able to ask a question in plain language can make that information easier to approach. It also creates a useful starting point for finding the documentation needed to understand the next step.

That fits the broader theme for me: making a complex platform easier to work with.

What I’m Taking Away

Customers need room to adopt Kubernetes and AI while continuing to run the applications they already have. The teams supporting them need a practical way to manage that growing responsibility.

What interests me most is how these capabilities can come together around a useful outcome: an employee finding the right answer, a developer getting an application into service, or a customer adopting a new technology with greater confidence.

That is where making complex things simple becomes meaningful.