Splunk Observability + Galileo: Bridging the AI Trust Gap

The era of the "agentic teammate" is no longer a futuristic concept— it’s here now. Acting on our behalf, AI agents are enabling us to reimagine key business workflows, streamline customer interactions, and automate daily tasks, so we can all focus on more strategic work.

Described by IDC as the “agentic pivot,” organizations will continue to incorporate agentic AI into every layer of the enterprise over the next couple of years. IDC predicts that “agentic automation will enhance the capabilities of more than 40% of enterprise applications by 2027, laying the foundation for next-generation AI operating models and reshaping one-third of business processes and workflows.”¹

However, agents are not simply “another service.” Unlike traditional software, agentic applications don’t always produce consistent outputs based on the same inputs or business logic. They behave and even fail in unpredictable ways that traditional approaches to observability weren’t designed to catch.

AI agents can hallucinate, take action without accurate context, and provide harmful responses. Not only can inaccurate agent behavior erode confidence with employees, but it can also be tremendously costly. A single agent interaction can consume thousands of tokens and rack up costs quickly. A harmful output can damage customer perception and negatively impact an organization’s bottom line.

That is the practical tension enterprises feel today — the need to use AI agents but the lack of confidence, governance, and guardrails to help ensure these agents are performing as intended. As AI agents take more ownership in driving mission-critical production environments, a new challenge has emerged: the AI trust gap.

Integrating Galileo Capabilities into Splunk Observability

This is why we are thrilled to welcome Galileo, its team, and technology to help close the AI trust gap for customers.

Cisco has completed its acquisition of Galileo, a company focused on evaluation, observability, and real-time protection for multi-agent systems across the agent development lifecycle. Its work aligns with today’s reality and what organizations need to thrive in the AI era — stronger evaluation mechanics, metrics that reflect agent quality and operational risk, and runtime controls that automatically intervene when agent behavior diverges from intent.

By integrating Galileo’s technology into Splunk Observability’s AI Agent Monitoring, we will supercharge the AI observability capabilities already available, so customers gain the confidence they need to help ensure agents’ performance.

Evaluation, Observability, Guardrails, and Controlled Token Costs

These are the three key areas where Galileo will help bridge the AI trust gap:

• Powerful and customizable out-of-the-box evaluations: We are providing powerful, custom and out-of-the-box AI evaluations that allow teams to visualize every agent, tool call and handoff alongside quality, cost and performance metrics. This provides visibility into how agents actually behave. Through access to a rich library of evaluators that span factual grounding, safety compliance, response consistency and more, teams can test, measure and improve agent prompts and outputs on day one.

• Real-time guardrails: Galileo doesn’t just detect undesirable behavior; it blocks it. By intercepting prompts and harmful outputs in under 200ms, we can help stop undesirable outputs like prompt attacks, hallucinations, and data leaks before they reach the end user.

• Cost-effective evaluations: By using Luna-2, Galileo's fast, low-cost family of small language models (SLMs) purpose-built for AI evaluation, we are enabling organizations always-on agent evaluations and guardrails at enterprise scale. Luna-2 models provide low latency and reduced costs for metric evaluations, so organizations don't need to rely on expensive LLM judges using frontier models. This provides continuous evaluation of production traffic and eliminates blind spots.

This gives AI engineering teams the ability to evaluate, observe, better protect, and improve AI agents across their development lifecycle.

The Future of Trustworthy AI

At Splunk, our mission has always been to provide visibility and insights that keep digital systems secure and reliable. And that mission must include the agents that power the businesses of today and tomorrow.

With its acquisition of Galileo, Cisco has helped solidify its position as an industry leader in AI observability. Together, Cisco, Splunk, and Galileo are helping customers move past the "trust gap" and toward a future where AI agents are not just powerful, but trustworthy and secure.

¹. IDC blog, “An IT tech leader’s guide through the agentic pivot,” November 20, 2025