Transforming observability data into intelligence with the agentic AI Assistant
AI assistants have become an essential part of the observability experience. Whether you're investigating service degradation, analyzing logs, understanding infrastructure health, or learning a new platform capability, you increasingly expect natural language interactions to help you get answers faster.
The first generation of AI assistants showed the power of generative AI. By combining large language models with observability data and platform tools, you could ask questions in plain English and get meaningful insights without deep expertise in query languages or product navigation. But this approach exposed a common challenge: consistency. Identical questions could result in different reasoning paths, tool selections, and ultimately different responses. Generative AI excels at reasoning and language understanding, but it can be less predictable when repeatedly running known operational workflows.
To solve this, Splunk Observability Cloud has evolved its AI Assistant with an agentic AI approach. Rather than relying solely on a language model to determine every step of an investigation, the Assistant now combines the reasoning capabilities of generative AI with purpose-built workflows designed for common observability tasks. The result is a more reliable, context-aware, and efficient assistant that helps you troubleshoot faster, reduce investigation time, and gain insights across your observability data with greater confidence.
This article walks you through what's changed, how agentic AI works, and what you can do with the latest AI Assistant capabilities.
How AI Assistant helps with this use case
Moving faster from insight to action
The latest version of AI Assistant, as of June 2026, is more than a feature upgrade — it's a step toward a more integrated and intelligent observability experience. By combining contextual awareness, broader knowledge support, flexible interaction models, and workflow continuity, AI Assistant helps you spend less time navigating complexity and more time resolving issues.
For you, that means faster investigations, less manual effort, easier knowledge sharing, and a more natural way to interact with observability data. Whether you're troubleshooting live incidents, learning the platform, or capturing findings for your team, AI Assistant helps turn observability insights into action more efficiently.
In conversations with one of our largest fintech customers, the impact of AI Assistant has been clear. By enabling application troubleshooting, increasing team self-sufficiency, and delivering significant time savings, AI Assistant helps both IT Operations and engineering teams spend less time on routine investigations and more time on innovation and product development.
Understanding why this matters to your team
For observability practitioners, the distinction between generative and agentic AI isn't about architecture. It's about outcomes. When you're investigating service degradation, application errors, or infrastructure issues, you need answers that are accurate, repeatable, and actionable. You need an assistant that can gather context, perform the right analysis, and provide guidance with confidence. As organizations increasingly rely on AI to accelerate troubleshooting and operational decision-making, consistency becomes just as important as intelligence.
This is where agentic AI changes the experience.
Combining intelligence with purpose-built workflows
The Splunk Observability Cloud AI Assistant has been upgraded from purely generative to agentic. Why? Because making an LLM figure out complex troubleshooting steps from scratch is slow and prone to errors. Now, the Assistant recognizes your end goal and automatically triggers optimized, battle-tested investigation workflows. The language model still acts as the natural interface to understand your intent, but the agentic system handles the heavy lifting, saving your team from manual orchestration fatigue.
This approach gives you several advantages:
- Improved consistency by following proven investigation patterns for common troubleshooting scenarios.
- Reduced hallucinations by grounding responses in validated workflows and platform data.
- Fewer follow-up questions because the Assistant can proactively gather the context required to complete an investigation.
- Faster time to resolution through automated running of multi-step analysis tasks.
- More comprehensive insights by correlating information across services, infrastructure, logs, traces, metrics, and user experience data.
The result is an AI Assistant that behaves less like a chatbot and more like an observability partner — one that can understand a goal, gather evidence across domains, and guide you toward answers with greater speed and confidence. For example, the screenshot below shows AI Assistant analyzing issues with the checkout service in the prod environment.
Exploring what's new in Splunk Observability Cloud AI Assistant
The Splunk Observability Cloud AI Assistant introduces a more powerful, intuitive, and context-aware experience designed to help you troubleshoot faster, explore observability data more naturally, and integrate AI-driven insights into your daily workflows. From generalized Q&A to context-aware responses, chat continuity, and workflow-friendly exports, AI Assistant is built to meet you where you are and help you move from question to action with less effort. Here's a closer look at what's new.
Getting quick answers with generalized Q&A
Splunk Observability AI Assistant now supports generalized Q&A for both Splunk Observability Cloud and observability concepts more broadly. In addition to answering product-related questions, AI Assistant can provide links to relevant documentation from the Splunk Observability Cloud docs library to help you go deeper. As shown below, AI Assistant provides guidance on detector creation, drawing directly from official Splunk Observability Cloud documentation.
How this helps you: You often need quick answers while you work — whether understanding a product's capability, clarifying an observability concept, or finding the right documentation. Instead of searching manually across docs, you can ask directly in the flow of work and get both an answer and supporting resources. This shortens time to information and makes AI Assistant useful beyond troubleshooting alone.
Grounding responses with automatic screen capture
AI Assistant is now context-aware and can automatically capture your current page view as context for the backend. It understands the view you're looking at and can use signals such as URL parameters and page context to ground its responses. In the example below, AI Assistant analyzes issues related to the services displayed on the page you're currently viewing.
How this helps you: This significantly reduces the need to manually describe where you are in the product or what entity you're investigating. Instead of writing long prompts to explain context, you can simply ask your question and let the Assistant infer relevant details from the page you're on. This makes interactions faster, more natural, and more accurate.
Investigating your way with an AI-native experience
Splunk Observability Cloud AI Assistant introduces a more flexible AI-native user experience. You can switch AI Assistant into full-screen mode for deeper investigations or float it alongside the product page for side-by-side analysis. The Assistant also supports more multimodal responses, including visualizations and charts. The screenshot below shows AI Assistant running in full-screen, AI-native mode.
How this helps you: Different workflows require different interaction models. Full-screen mode gives you more space for deeper analysis, while floating mode lets you stay anchored in the product experience as you investigate. Multimodal responses make complex information easier to understand at a glance, helping you interpret trends, anomalies, and system behavior more efficiently.
Saving and sharing insights with PDF exports
You can now export AI Assistant conversations as PDFs, making it easy to save prompts and responses for use in existing workflows. The example below shows a chat conversation exported as a PDF.
How this helps you: Investigations often need to be shared or documented after the fact. PDF exports let you quickly preserve troubleshooting context and reuse it in incident postmortems, internal documentation, Confluence pages, runbooks, and team knowledge bases. This helps your team operationalize insights from AI Assistant instead of losing them when the conversation ends.
Picking up where you left off with chat history
Splunk Observability Cloud AI Assistant now includes chat history, letting you switch between conversations, reopen previous threads, and continue investigations seamlessly. Each thread includes an LLM-generated title to help you quickly identify the topic. You can also rename or delete threads as needed. As shown below, AI Assistant displays historical chat conversations, allowing you to resume prior conversations.
How this helps you: Troubleshooting is often iterative and spreads across time. Chat history makes it easier to revisit previous investigations without starting from scratch. Auto-generated titles improve discoverability, while rename and delete controls give you flexibility in organizing your conversations. This creates a more persistent and practical AI Assistant experience for day-to-day use.
Moving forward with smart prompt suggestions
AI Assistant now provides smart prompt suggestions throughout the Splunk Observability Cloud experience. These suggestions are tailored to the page you're on, and after a conversation begins, AI Assistant continues recommending prompts that can help you move forward in your investigation. In the example below, AI Assistant surfaces page-specific prompts to make getting started easier.
How this helps you: Many users know what they want to accomplish but might not know the best way to ask for it. Smart prompt suggestions reduce that friction by guiding you toward relevant, high-value interactions. They also help you discover new ways to use AI Assistant, making it easier to get started and maintain momentum during investigations.
Reviewing supported capabilities
AI Assistant helps you quickly understand and troubleshoot your applications and infrastructure across the full observability stack. It provides natural-language access to application performance, infrastructure health, logs, alerts and incidents, end-user experience, synthetic monitoring, metrics analysis, and advanced analytics — enabling faster investigation, root cause analysis, and operational insights without requiring deep expertise in underlying query languages or data models. The following areas are covered by AI Assistant.
| Product Area | Description |
|---|---|
| APM | Application performance — services, traces, error rates, endpoints, and latency across your distributed system |
| Infra | Health of hosts, Kubernetes nodes & pods, and resource utilization |
| Logs | Log data tied to your services and infrastructure — supports both service and unified identity types |
| Alerts/Incidents | Search alerts/incidents related to an entity, environment, ID, or other keywords |
| RUM | Real end-user experience — app health, crashes, Core Web Vitals, sessions (BRUM & MRUM) |
| Synthetics | Provides details on Synthetics test, test run details, downtime configurations, private locations, and information present in the UI |
| Metrics Usage | Visibility into metrics consumption — cost optimization and MTS cardinality spike analysis |
| Metrics Lookup | Search metric names, metric families, dimensions, properties in natural language |
| SignalFlow | Splunk's analytics query language — generated for you from natural language |
Looking ahead
The future of observability isn't just about having more data—it's about making that data instantly meaningful.
As we continue to evolve the Splunk Observability Cloud AI Assistant, our vision is to create an experience that moves beyond reactive question-and-answer interactions to become a more proactive partner in your workflow. We believe AI should do more than wait for prompts. It should understand where you are, recognize what you're trying to accomplish, and surface relevant insights at the right moment.
That means building an AI assistant that's increasingly aware of both user interactions and system context across the observability platform. As you move through your workflows, AI Assistant can help by automatically highlighting what might need attention, bringing forward relevant context, and guiding you toward the next best action. Instead of spending valuable time piecing together signals or figuring out what to ask, you can focus more directly on understanding impact, accelerating decisions, and resolving issues.
This is where AI-powered observability creates real value: reducing the effort required to extract insight from complex telemetry and helping teams move faster from detection to understanding to action. Whether identifying patterns across domains, surfacing investigation context earlier, or helping you navigate incidents more efficiently, AI Assistant can become a force multiplier for engineering and operations teams.
Our long-term vision is to make AI Assistant a deeply integrated part of the observability experience—one that helps you stay ahead of issues, work more efficiently, and focus on business outcomes rather than manual analysis. In that future, AI is not just an interface layer on top of observability data; it's an active participant in helping teams unlock the full value of their systems and respond with greater speed and confidence.
Additional resources
These resources might help you understand and implement this guidance:
- Splunk Resource: AI Assistant in Observability Cloud
- Splunk Help: Splunk AI Assistant in Observability Cloud