Enable customers to use an AI-powered chat interface in the OpenShift console to understand and optimize cloud costs through natural language queries and guided analysis.

Nature and Description of Request

Customers need an accessible, conversational interface for understanding cloud costs and discovering optimization opportunities in their OpenShift clusters. This feature is part of the broader "Applied AI" initiative that uses AI as part of the implementation (via LLMs and MCP servers) to enhance the user experience, complementing the existing troubleshooting experience (OCPSTRAT-2608).

Current Limitation

Today, customers cannot use a chat interface to ask open-ended questions about their cluster's cloud costs or optimization opportunities. While OpenShift provides powerful cost optimization tools (cluster autoscaler, VPA, HPA, spot instances, instance sizing), customers must:

• Manually discover and configure these tools
• Understand complex Kubernetes resource metrics
• Navigate multiple console views and documentation
• Have deep knowledge of cloud provider pricing models

This creates friction that leads to:

• User abandonment before achieving cost optimization
• Underutilization of existing cost optimization features
• Higher operational overhead for cluster administrators
• Missed cost savings due to complexity barriers

Desired Behavior

• Customers should be able to:
• Open a chat interface* in the OpenShift console
• Ask natural language questions* about cost optimization, such as:
  • "What are my top 3 cost optimization opportunities?"
  • "Why is my cluster costing more this month?"
  • "How can I reduce costs for my development clusters?"
  • "What instance types should I be using?"
• Receive guided analysis* through a specialized prompt that walks them through cost analysis
• Get actionable recommendations* pointing to existing tools:
  • Cluster autoscaler configuration
  • Vertical Pod Autoscaler (VPA) setup
  • Horizontal Pod Autoscaler (HPA) recommendations
  • Spot instance opportunities
  • Instance size optimization
• Query cluster state* via Kubernetes/OpenShift MCP servers to provide context-aware recommendations
• Access read-only analysis* (Phase 1 scope - no automated changes to cluster state)

Use Case

Typical customer workflow:

• Customer opens OpenShift console and navigates to Applied AI chat interface
• Asks: "How can I optimize costs for this cluster?"
• AI loads cost optimization analysis prompt
• AI queries cluster state via MCP servers (node pools, instance types, resource utilization, pod configurations)
• AI analyzes data and provides:
• Current cost drivers (e.g., "40% of your costs are from overprovisioned node pools")
• Specific recommendations (e.g., "Enable cluster autoscaler to reduce idle capacity")
• Links to relevant tools and documentation
• Customer follows recommendations to configure existing cost optimization features
• Customer returns to chat to ask follow-up questions or explore additional optimizations

Target platforms:

Self-managed OpenShift on AWS, Azure, and GCP.

Out of scope:

• ROSA/ARO managed services
• Other clouds
• On-prem 

Note: we could start with AWS if that is easier.

Phase 1 scope

Read-only analysis and recommendations.

Out of scope:

• cost projections
• automated remediation

Business Requirements

Customer Impact

Primary segment: Small and medium businesses (SMBs) who need accessible cost optimization tools

Customer requests: SMBs have requested easier access to cost optimization capabilities without requiring deep Kubernetes expertise

User experience improvement: This feature provides an accessible entry point for cost optimization that complements the existing suite of tools

Strategic Value

Applied AI initiative: This is a key component of OpenShift's comprehensive "Applied AI" suite, which includes:

• Troubleshooting assistance (OCPSTRAT-2608)
• Cost optimization (this feature request)
• Additional AI-powered operational experiences (planned)

Strategic objective: Lower Total Cost of Ownership (TCO) for OpenShift clusters by:
Reducing operational overhead* through improved user experience

• Decreasing friction* in adopting cost optimization tools
• Improving accessibility* for users without deep Kubernetes expertis
• Demonstrating value* through cost savings facilitation

Business Justification

Without this capability:

• Higher friction* leads to user abandonment before achieving cost optimization
• Underutilization* of existing cost optimization features (autoscaler, VPA, HPA)
• Increased support burden* from customers struggling with cost management
• Missed opportunity* to differentiate OpenShift with AI-powered experiences
• Competitive disadvantage* as other platforms invest in AI-assisted operations

With this capability:

• Lower barrier to entry* for cost optimization makes tools more accessible
• Improved user satisfaction* through intuitive, conversational interface
• Higher adoption* of existing cost optimization features
• Reduced operational costs* for customers (supporting TCO reduction value proposition)
• Strategic positioning* as part of comprehensive Applied AI suite

Timing

Why now:

• Applied AI initiative is actively under development (OCPSTRAT-2608 and related work)
• Creating a comprehensive suite of AI experiences requires multiple use cases
• Cost optimization is universally relevant and complements troubleshooting capabilities
• Market timing: AI-powered operational tools are becoming competitive differentiators

Affected Packages and Components

Teams

Primary teams:

• Autoscaling Team*: Integration with cluster autoscaler, VPA, HPA; cost optimization recommendations

Other teams

• Console Team: I think there is no necessary work on console side
• Node/MCO/other: other teams could take this work on, it should be following a pretty clear pattern for applied AI UXs.

Technical Components

Core components:

• openshift-console*: Chat interface UI, Applied AI integration point
• cluster-autoscaler-operator*:s Autoscaling recommendations and configuration
• vertical-pod-autoscaler-operator*: VPA recommendations
• horizontal-pod-autoscaler*: HPA recommendations
• MCP servers*: Kubernetes/OpenShift MCP servers for querying cluster state

Related Services

• MCP/Applied AI platform: LLM hosting, MCP server orchestration, prompt library
• Cost monitoring backend: Cloud provider billing data aggregation
• Documentation integration: Links to autoscaler, VPA, HPA setup guides

Additional Context

Related work:

• OCPSTRAT-2608: Applied AI troubleshooting experience
• Broader Applied AI initiative for operational UX improvements

Out of scope for Phase 1:

• Cost projections ("You'll save $X/month")
• Automated remediation (applying changes without user approval)
• ROSA/ARO managed service integration (focused on self-managed OpenShift)