Cisco enhances Splunk observability for AI token usage tracking

Cisco has expanded Splunk Observability to offer visibility into enterprise AI usage, leveraging capabilities from its recent Galileo acquisition. The enhanced features allow for tracking tokenomics, costs, and productivity related to AI coding tools, with further coverage provided by SDx Central.

The updated Splunk Agent Observability now provides tokenomics tracking within Splunk Observability Cloud and Cisco Cloud Control. This enables users to monitor the usage, costs, and productivity of AI coding tools like Claude Code and Codex, aggregating spend across different vendors into a single view. Cisco's Deep Time Series Model (CDTSM) tool also offers predictions for final billing before the end of a cycle. Kamal Hathi, SVP and GM for Splunk, highlighted the growing importance of AI costs for customers, stating the new solution aims to optimize AI token spend, manage consumption, and actively control costs.

This integration aligns token usage with business priorities, providing full visibility and attribution, along with prescriptive tools for optimization without compromising quality. The move into tokenomics is spearheaded by Splunk, building on Cisco's earlier discussions about its Galileo acquisition, which aids in analyzing AI agents at a behavioral level. Splunk's approach layers model and GPU assessment to evaluate performance and precisely count tokens for AI actions.

The integration aims for minimal intrusion, with low-lift integration into third-party systems that consume significant tokens. This enhanced observability also extends to sovereign AI, allowing for better tracking and cost comparisons between local and cloud infrastructure.

Source: SDx Central