Splunk Blogs

Learn to diagnose the six primary hallucination patterns and implement a robust pipeline of retrieval, reranking, and guardrail enforcement to ensure reliability.

Splunk's new add-ons let teams track who used AI tools, what changed, and whether API keys were created or misused across ChatGPT, Claude, and other providers.

These .conf26 sessions will guide practitioners through building ML-based detections, understanding AI agent risks, and applying these techniques to real attack scenarios.

Learn the eval-to-guardrail lifecycle to transform offline evaluation criteria into runtime policies that intercept and block agent failures in production.

Use real-time output guardrails and remediation strategies to surgically contain AI agent failures before they reach your users.

Learn how custom AI evaluation metrics outperform standard benchmarks to ensure GenAI reliability, compliance, and business alignment for confident deployment.

In August, the Splunk Threat Research Team had 2 releases of new security content.

Learn how to maintain production RAG performance by monitoring filter selectivity, tuning hybrid search parameters, and tracking retrieval drift.

Splunk now supports Google Cloud Private Service Connect.