Cisco Empowers Organizations to Govern AI at Scale with Full-Stack Observability

To gain confidence and trust in AI, today’s enterprises need deep visibility into their AI tech stack and insights to govern the performance, cost, quality, and security of their agentic AI investments. Agentic AI can be unpredictable—agents can hallucinate, provide harmful responses, and be incredibly costly. A single agent interaction can consume runway costs and resources.

At .conf26, Cisco is unveiling new Splunk Observability innovations to provide a unified, full-stack perspective that evaluates, observes and controls agent behavior, and correlates agent performance across the entire AI stack, down to hardware. Through these innovations, organizations can gain a unified view into both their infrastructure health and the performance of their AI agents to optimize AI at scale.

In addition, Cisco is announcing new innovations to help organizations build more resilient infrastructure and eliminate visibility gaps and cost barriers so teams can troubleshoot with confidence and efficiency.

Bridge the AI Trust Gap by Evaluating, Observing, and Controlling AI Agents

Organizations need insights to evaluate, observe, and protect AI agents across their development lifecycle to use AI with confidence. Through new Splunk Observability innovations, Cisco is enabling enterprises to ensure AI agents are performing as intended:

• Now offered in Splunk Observability Cloud and in Cisco Cloud Control, Splunk Agent Observability evaluates agent behavior, observes performance across the entire AI stack, enables token usage and cost optimization, and controls agents in real-time with runtime guardrails to block harmful actions and outputs for every response. Through these insights, organizations can transform AI use from an unpredictable expense to an accelerator for growth.

Optimizing Tokenomics: Visualize AI Token Spend, Quality, and Infrastructure in One View

To manage spiraling AI costs, organizations must embrace AI tokenomics as a core business discipline. AI tokenomics is the practice of tracking and optimizing token consumption and usage, then tying that spend to business outcomes.

Today, Cisco is introducing a dedicated Tokenomics capability to track and attribute AI token expenditure in real-time, helping to operationalize a tokenomics framework:

• Splunk Agent Observability’s new Tokenomics capability surfaces coding agent usage, cost, adoption, and productivity insights at the team and user level. Splunk provides one place to view the cost of employee usage of Claude Code, Codex, and other AI coding tools. It also consolidates consumption across providers into one view and will forecast where that spend is heading before the billing period ends, using the Cisco Deep Time Series Model (CDTSM).

By providing deep and accurate insights into token usage and trends, Splunk Agent Observability enables organizations to make data-driven budgeting or optimization decisions, e.g. to route workloads to the most cost-effective model. The new capability is offered in Cisco Cloud Control and in Splunk Agent Observability on-premises and in Splunk Observability Cloud.

Maximize Infrastructure Resilience and Troubleshoot Autonomously

Cisco is helping organizations strengthen their infrastructure resilience by helping to eliminate visibility gaps and cost barriers that can hinder autonomous troubleshooting. Through these new innovations, teams can troubleshoot autonomously and scale with confidence:

• Splunk Observability Studio solves the problem that new apps are often built and deployed before teams understand how to monitor, debug, or operate them. Through Observability Studio, teams can make sure every new application is “born observable,” or measurable and operationally ready, using OpenTelemetry-compliant instrumentation — from day one.

• Network Intelligence App brings Cisco network topology, device health, and events into Splunk, so network teams can trace an alert straight to the device behind it and the network around it without leaving Splunk.

• Splunk Observability Cloud Free Edition provides access to a free tier of full-stack observability for small environments. This edition enables teams to experience the platform’s capabilities in real workflows—without upfront cost or procurement barriers or time-bound trials.

• Simplified Editions for Observability Cloud align Splunk’s offerings with the evolving observability landscape. The Essentials edition focuses on core observability use cases designed for teams getting started, while the Premier edition is for teams who are rapidly building agents at scale and need more capabilities. Both editions include Splunk Agent Observability, APM, Infrastructure Monitoring, and our new Observability Logs is included in both editions, providing developers and ITOs teams with cost-effective log analytics entirely within Splunk Observability Cloud.

The Path Forward

By unifying agent monitoring, tokenomics and infrastructure health into a unified full-stack AI observability experience, Cisco is empowering the global enterprise to optimize AI at scale with confidence and precision.