WORKSHOP

Build Better, Spend Less: A Hands-On Workshop on Token Efficiency

As AI agents move from experimental prototypes to critical business workflows, the "black box" of token consumption has become a primary concern for engineering teams. Unchecked token usage doesn't just inflate cloud bills; it obscures the relationship between infrastructure spend and actual model performance. To truly scale, practitioners need more than just high-level dashboards, they need the ability to correlate granular tokenomics with specific agent behaviors and business outcomes.

In this 90-minute hands-on session, we move beyond the basics to help you take full control of your AI stack. You will work directly with a multi-agent travel planner application, gaining live experience in instrumenting workflows with Splunk Agent Observability (powered by the acquisition of Galileo). By the end of the session, you will have the practical skills to visualize usage patterns, conduct side-by-side efficiency comparisons, and ensure your deployments are as cost-effective as they are intelligent.

What you will learn:

• Rapid Instrumentation: Learn the fastest path to integrating Galileo into your existing agentic workflows to gain immediate, high-fidelity visibility into your infrastructure
• Granular Token Visibility: Gain precise oversight of token consumption across all dimensions, allowing you to pinpoint resource allocation and eliminate hidden inefficiencies in your agent workflows.
• Performance Bake-offs: Conduct comparative testing between different configurations to identify the optimal balance of performance and token efficiency for your specific requirements.
• Strategic Optimization with Luna: Drastically reduce token costs and latency by deploying Luna models for evaluation. This approach enables you to maintain 100% traffic visibility and high accuracy, driving massive annual savings while scaling your agent workflows with confidence.