AWS re:Invent 2025 - Using GenAI to profile, scale, and optimize your multi-tenant SaaS architecture

Leveraging GenAI to Optimize SaaS Architecture and Operations

Overview

• Presentation covers how to use Generative AI (GenAI) to profile, scale, and optimize multi-tenant SaaS architecture
• Presented by AWS Solutions Architects Dave Roberts and Damakashrian
• Explores key areas where GenAI can drive SaaS maturity and business impact

Challenges of Growing a SaaS Business

• Increasing customer base for more revenue
• Driving more revenue per customer
• Reducing operating costs and improving efficiency

Best-in-Class SaaS Vendors

• Leverage data to understand cost to serve customers
• Gain insights into customer behavior and usage patterns
• Focus on resource efficiency to scale cost savings
• Use data-driven insights to create optimal packaging and pricing

Incorporating GenAI into the SaaS Control Plane

• Expose control plane capabilities (metrics, tenant data, workflows) as MCP servers
• Use GenAI agents to query data, derive insights, and automate workflows
• Establish a modular, extensible architecture to scale GenAI capabilities

Cost Analysis and Optimization

• Leverage GenAI agents to:
  • Attribute infrastructure costs to individual tenants
  • Calculate accurate cost per tenant metrics
  • Analyze historical cost and margin trends
  • Forecast future cost and margin projections
• Integrate cost insights into SaaS admin dashboards
• Use GenAI to recommend pricing and packaging optimizations

Customer Insights and Behavior Analysis

• Leverage access logs and application instrumentation to capture feature usage data
• Use GenAI to identify customer personas based on usage patterns
• Analyze persona profitability to inform targeted marketing and packaging
• Automate customer success workflows (churn prevention, upsell) using GenAI

Resource Optimization and Efficiency

• Implement GenAI-powered anomaly detection to identify potential issues
• Use pattern recognition to classify critical conditions (noisy neighbors, resource misuse)
• Orchestrate automated remediation workflows with human-in-the-loop verification
• Build a knowledge base of historical issues and resolutions to improve decision-making

Key Takeaways

• Start small, but build an extensible GenAI-powered control plane architecture
• Leverage pre-built AI agents and MCP servers to quickly derive business-critical insights
• Integrate GenAI into core SaaS operations to drive efficiency, profitability, and growth
• Empower business stakeholders with natural language interfaces to the GenAI capabilities
• Continuously expand GenAI's role as the SaaS matures and new requirements emerge