> ## Documentation Index > Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt > Use this file to discover all available pages before exploring further. # SkyLine Air Kubernetes Implementation > Framing a FinOps AI agent to automate cloud cost management, correlate billing and Kubernetes telemetry, and guide SREs deploying and operating the agent on AWS EKS Welcome back. In this lesson we’ll pause to establish context before building the FinOps AI agent. Understanding the real-world problem and where an agent fits inside an engineering organization will make the implementation decisions clearer and more practical. Goals for this lesson: * Explain why we need a FinOps AI agent and the problems it solves. * Show where such an agent sits in a typical engineering organization and infrastructure stack. * Set expectations for the architecture and implementation steps that follow. This lesson focuses on problem framing and placement of the FinOps AI agent. Later lessons will cover architecture, data pipelines, model selection, and deployment on Kubernetes (EKS). Why "FinOps AI agent"? In short: to automate and scale cloud cost management, surface actionable cost insights, and assist engineers with contextual, prioritized recommendations that reduce waste without blocking delivery. Meet Elena. Elena is a senior DevOps engineer who has just joined SkyLine Air — a European airline operator that serves about 50 destinations. Even modest airlines run many systems: booking, auditing, operations, legal, and more. Elena’s day-to-day will involve keeping these systems performant and cost-efficient. ![A presentation slide titled "Why a FinOps AI Agent?" with the heading "Why We're Building This" and a smaller "FinOps AI Agent" label. Below the text is a rounded square icon showing a simple robot face.](https://mintcdn.com/kodekloud-c4ac6d9a/rjGeTDj8cr6D5YbW/images/Building-Kubernetes-FinOps-AI-Agent/The-Problem-Statement/SkyLine-Air-Kubernetes-Implementation/finops-ai-agent-slide-robot-icon.jpg?fit=max\&auto=format\&n=rjGeTDj8cr6D5YbW\&q=85\&s=3dcbdfadce0794650003e0b474f38eef) SkyLine Air runs a broad set of services and internal tools. The booking and auditing systems are especially critical; they require reliability and have strict latency and availability SLAs. Other systems such as operations tools and back-office services are equally important but have different cost and scaling profiles. ![A slide titled "A Lot to Run" showing three square icons labeled Booking, Auditing, and Operations. The Booking and Auditing icons are highlighted in dark blue (tag and magnifier), while Operations is greyed out with a wrench.](https://mintcdn.com/kodekloud-c4ac6d9a/rjGeTDj8cr6D5YbW/images/Building-Kubernetes-FinOps-AI-Agent/The-Problem-Statement/SkyLine-Air-Kubernetes-Implementation/booking-auditing-highlighted-operations-greyed.jpg?fit=max\&auto=format\&n=rjGeTDj8cr6D5YbW\&q=85\&s=228daf3036c3833a8a4e665f7f186aee) By the numbers, SkyLine Air is small-to-medium but with meaningful cloud spend. The company has roughly 500 employees, about 200 of whom are in tech and product — enough to generate significant cloud usage and complexity. ![An infographic titled "By the Numbers" on a dark background with a circular progress chart showing "500 employees." To the right it highlights "200 Tech and Product" staff.](https://mintcdn.com/kodekloud-c4ac6d9a/rjGeTDj8cr6D5YbW/images/Building-Kubernetes-FinOps-AI-Agent/The-Problem-Statement/SkyLine-Air-Kubernetes-Implementation/by-the-numbers-500-employees-chart.jpg?fit=max\&auto=format\&n=rjGeTDj8cr6D5YbW\&q=85\&s=35d839da4c21f8427ec903b5eadae5ab) Elena joins the Site Reliability Engineering (SRE) team. This team manages a large Kubernetes cluster hosted on AWS EKS. The cluster runs the customer-facing web UI, booking services, backend APIs, and other critical workloads — making it a central place to observe and act on cost and performance signals. ![A simple diagram titled "Everything Runs Here" showing UI, App, and Backend icons all connected to an AWS EKS (Kubernetes) cluster.](https://mintcdn.com/kodekloud-c4ac6d9a/rjGeTDj8cr6D5YbW/images/Building-Kubernetes-FinOps-AI-Agent/The-Problem-Statement/SkyLine-Air-Kubernetes-Implementation/ui-app-backend-eks-diagram.jpg?fit=max\&auto=format\&n=rjGeTDj8cr6D5YbW\&q=85\&s=265a2446fa926a2f9662616bcc914900) Key operational and cost challenges Elena will face * Lack of cross-team visibility into cloud spend and resource waste. * Difficulty prioritizing cost-saving actions vs. reliability and performance. * Time-consuming manual investigations to identify overprovisioned resources or idle workloads. * Multiple data sources (billing, cluster metrics, deployment manifests) that require correlation. How a FinOps AI agent helps * Correlates billing and usage data with Kubernetes telemetry. * Proactively surfaces high-impact recommendations (right-sizing, idle pod detection, savings from reserved instances). * Provides contextual guidance to SREs and developers so cost actions are safe and prioritized. * Automates repetitive diagnostics and creates tickets or PRs for low-risk changes. Quick reference: useful links * [FinOps Foundation — What is FinOps?](https://www.finops.org/) * [Kubernetes Documentation](https://kubernetes.io/docs/) * [AWS EKS Documentation](https://docs.aws.amazon.com/eks/latest/userguide/what-is-eks.html) Next: we will translate this problem framing into concrete requirements for our FinOps AI agent — data sources, decision logic, and the architecture for deploying the agent into the EKS cluster. > **[Watch Video](https://learn.kodekloud.com/user/courses/building-kubernetes-finops-ai-agent/module/5d434b76-bd5d-4526-b9c6-4a994c195994/lesson/0f2da9da-06d8-453a-b401-fa64e4f96291)** This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.