{"slug": "what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs", "title": "What's Next in Agentic Operations: Introducing AI Week at Grafana Labs", "summary": "Grafana Labs announced AI Week, a series of daily announcements about new AI capabilities in Grafana Assistant, the agentic LLM in Grafana Cloud, and a suite of AI tools for testing code, investigating incidents, evaluating agents, and automating operations. The company aims to help users integrate observability before deployment as agentic operations increase the rate of change and complexity. AI Week runs weekdays with technical demos, blog posts, and Reddit Q&A.", "body_md": "# Explore what's next in agentic operations: Introducing AI Week\n\nObservability has traditionally been tacked on after your code hits production, but with agentic operations on the rise, that's no longer sustainable.\n\nAgents have dramatically increased the rate of change as they write more code, ship more changes, and operate more systems—all at a speed that compounds scale and complexity. As a result, observability needs to be factored in *before* deployment, with instrumentation written alongside the code, agents evaluated before they ship, and plans informed by production data.\n\nAI Week is designed to help you navigate that shift. We’re dropping big announcements every day about new AI capabilities in __ Grafana Assistant__, the agentic LLM in Grafana Cloud, and our growing suite of AI tools that will help you test your code before you ship as well as investigate incidents, evaluate agents, and automate routine operational work.\n\nThroughout the week you'll find:\n\n- Technical demos on\n[YouTube](https://www.youtube.com/grafana) - Specific use cases to experiment with on\n[our blog](https://grafana.com/blog/?pg=explore-what-s-next-in-agentic-operations-introducing-ai-week&plcmt=in-text) - Behind-the-scenes looks at how we build our tools on our new engineering blog,\n[Unprompted](https://medium.com/grafana-labs) - Our AI team ready to chat and answer questions on\n[Reddit](https://www.reddit.com/r/grafana/)\n\nThe goal is to support you wherever you are on your AI journey. Whether you're preparing your first AI-powered application for production or coordinating agentic workflows across your environment, the AI capabilities in Grafana Cloud can help prepare software for production so it's observable, understandable, and ready to operate.\n\nOn the flip side, if you're curious (or nervous) about how your agents are performing, you can tap into Agent Observability, the solution in Grafana Cloud that provides high-level insights into agent behavior. But it doesn't just observe. Built-in evaluations continuously score the quality of your agents' responses, while guards run in real time to enforce policies and block unwanted behavior, catching hallucinations, drift, or policy violations before your customers do.\n\nWe're excited to show you what's new in AI for observability and observability for AI—and for you to get hands on with all the latest AI tools during our first ever AI Week.\n\nBig AI announcements on all your favorite weekdays\n\nEach day we'll focus on a different part of how AI is changing operations and observability. To help you prepare, here's a peek at what we're making generally available, a preview of some exciting new features, and how you can engage with the AI team to learn more.\n\nOne of the highlights of AI Week is also how easy it is to get started and try all the AI tools that roll out this week.\n\n**If you're a Grafana Cloud user**, you'll have immediate access to all the AI capabilities in Grafana Cloud. If you don't have an account,which comes with generous limits for all our offerings.[sign up today for our free plan,](https://grafana.com/auth/sign-up/create-user/?pg=explore-what-s-next-in-agentic-operations-introducing-ai-week&plcmt=in-text)\n\n**If you're a Grafana Enterprise or Grafana OSS user,** you can access many of these AI features by creating a Grafana Cloud account and connecting it to your Grafana installation.[via a one-click setup](https://grafana.com/docs/grafana-cloud/machine-learning/assistant/on-premise/?pg=grafana-assistant-everywhere&plcmt=in-text)\n\n**If you're new to the Grafana-verse, welcome!** You can try out, no registration required. This sandbox for testing and learning about features in Grafana Cloud allows you to jump right into Grafana Assistant or any of the solutions on our open observability platform before signing up for an account.[Grafana Assistant on Grafana Play](https://play.grafana.org/d/to6j8mh/grafana-play-home)\n\nAdd observability from the very start\n\nTo kick things off, Monday's focus will be on those early stages of onboarding and planning, making sure observability is there right from the start. We'll be talking about the new immersive Grafana Assistant Workspace feature, which gives you the best experience for interacting with Assistant.\n\nYou can ask questions of telemetry in natural language, which means you can understand unfamiliar systems and onboard faster than ever. And when it comes to decisions and planning changes, agents can do deep research with your production telemetry to inform plans and make decisions.\n\nWhat's new on Monday\n\n**Assistant Workspace: **Now generally available, __ Workspace__ gives you a home for your team's conversations from the very start. You get chat history, a live canvas, and investigation reports in one place—no need to do everything in a sidebar. And if you pair it with Grafana Assistant Investigations, a running investigation becomes a shareable report without you having to worry about any rewriting.\n\nMove from planning to acting\n\nOnce your plan is in place, you need to start building. Today's fast moving engineering creates new observability problems, so Tuesday will focus on how you can use AI tools and your existing telemetry to feel confident that your code is properly instrumented and optimized for your goals. We'll also discuss how you can use our gcx command line tool to build improvement loops, all informed by the realities of production.\n\nWhat's new on Tuesday\n\n**gcx:** __ Grafana's agentic CLI__ for managing dashboards, alert rules, data sources, and other resources as code will become generally available. gcx brings self-managed Grafana and Grafana Cloud under one command line, with agent-friendly input and output for tools like Claude Code, GitHub Copilot, and Cursor, and full GitOps support for versioning dashboards and alerts in Git.\n\n**Grafana MCP server:** Give any MCP-compatible client, including Claude Desktop, Codex, Cursor, and more, direct access to a Grafana instance's dashboards, alert rules, incidents, and data sources. No more pasting metrics into a chat window; the agent queries live telemetry from where the code is being written. __ Grafana MCP server__ will be generally available for Grafana OSS and Grafana Cloud.\n\nTrack and resolve incidents faster\n\nAgents are driving higher traffic to our services as people automate more workflows. This can mean more alerts, more questions, and more stressful incidents. On Wednesday, we’ll show how to get answers faster than ever when things don’t look right.\n\nWe'll show you how to use our AI so the next time an alert fires, you'll get a vetted set of hypotheses and recommendations based on analysis that follows industry best practices. By the time you open Grafana Cloud, you'll have everything you need to act swiftly.\n\nWe'll also show how to get regular, automated insights into the health of your system, and we'll give you a sneak peek at a new feature that brings agentic intelligence to your alerts.\n\nWhat's new on Wednesday\n\n**Assistant Investigations:** Generally available as of Wednesday, __ Assistant Investigations__ forms hypotheses and swarms over the problem to chase down the leads in the data, proving or disproving each one. You can stay in the driver's seat and steer the investigation, or step back and wait a few minutes and let Assistant Investigations finish the task and hand back a conclusion.\n\n**Assistant Automations: **Also generally available on Wednesday, __ Automations__ let a saved Grafana Assistant prompt run again automatically, on a schedule or on demand, so recurring checks like a daily error rate summary are sent to your Slack channel without anyone re-typing the same question every morning.\n\n**Watcher Agents:** Moving into public preview the same day, Watcher Agents are built for the things you can't write an alert rule for. You can't always spell out every way a system might go wrong, but you know what right looks like. So just describe that: Tell the watcher agent what to monitor, point it at the right data sources, dial in its sensitivity, and share the operational context it can't infer on its own. From there, the agent calibrates its own queries and baselines and looks for the moment right starts turning into something very wrong in your system.\n\nMaintain trust in the age of agentic tooling\n\nYour organization is probably building all sorts of agents these days, and you need to confidently make changes without worrying if it will impact performance at scale. We faced a similar obstacle when Assistant grew faster than expected last year, so we know this challenge all too well.\n\nThursday will be dedicated to showing you how to adapt to this new world. Specifically, we'll show you how we took the most useful lessons from that experience and packaged it into a solution for iterating on agents. And speaking of iterating on agents, we can’t wait to share __ how we’re using agent arenas__ to pit iteration against iteration to evaluate which version wins.\n\nWhat's new on Thursday\n\n**Agent Observability: **Generally available on Thursday, __ Agent Observability__ extends Grafana Cloud's OpenTelemetry-native monitoring to the AI systems teams are now shipping. It watches your agents, tests their behavior, and steps in when something's off. Instrumented agents emit telemetry that captures the usual signals (usage, latency, and errors) alongside newer signals that matter specifically for AI systems: token usage/cost and the conversation itself. Conversation data gives teams forensic-level debugging down to the individual exchange. Built-in evaluators go a step further, continuously scoring conversations to check agent behavior directly while guards enforce your policies in real time, blocking unwanted behavior, such as hallucinations, drift, or policy violation, before a customer does.\n\nStay in touch with teammates and your agents\n\nWith so much speed and so many moving parts, collaboration has never been more important than it is today. But not every team or organization communicates in the same way, which is why we've always believed in meeting teams where they are. We'll close the week discussing how we're carrying that same philosophy over to the world of agentic operations, and how you can make sure everyone's in the loop, whether it's on-premises or in the cloud, mobile or desktop, Slack or Teams.\n\nWhat's new\n\n**Assistant support for Microsoft Teams:** Available in private preview on Friday, you can access observability intelligence right in the chat where anybody can bring insights into group discussions. This new integration follows in the footsteps of our Slack integration, which is generally available.\n\n**Assistant on mobile: **With our upgraded Grafana Mobile app we're putting Assistant in your pocket. Assistant on mobile will roll out in public preview so you can keep things running while you’re on the run.\n\nHow to stay up to date and in touch all week\n\nThroughout AI Week, the AI team will be eager to chat on Reddit or on our social media platforms to answer questions and learn more about your AI use cases. We'll also share updates about the features discussed here (and more!) as they happen.\n\nAlso don't forget to sign up for our upcoming webinar [\"How to monitor AI agents in production with Grafana Agent Observability\"](https://grafana.com/go/webinar/how-to-monitor-ai-agents-in-production-with-grafana-ai-observability/?pg=explore-what-s-next-in-agentic-operations-introducing-ai-week&plcmt=in-text) on Aug. 12. You'll see demos of conversations, traces, token cost, and live quality checks for AI agents correlated with your apps and infrastructure in Grafana Cloud.\n\nHead over to __ Grafana.ai__ to learn all the details about our AI offerings, and thanks for joining us this week!\n\n[Grafana Assistant](/products/cloud/ai-assistant//?pg=explore-what-s-next-in-agentic-operations-introducing-ai-week&plcmt=footer-cta)* is the easiest way to get started with metrics, logs, traces, dashboards, and more in Grafana Cloud. We have a generous forever-free tier and plans for every use case. *[Sign up for free now!](/auth/sign-up/create-user/?pg=explore-what-s-next-in-agentic-operations-introducing-ai-week&plcmt=footer-cta)\n\nTags", "url": "https://wpnews.pro/news/what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs", "canonical_source": "https://grafana.com/blog/explore-what-s-next-in-agentic-operations-introducing-ai-week/", "published_at": "2026-07-27 12:38:48+00:00", "updated_at": "2026-07-27 12:52:26.648104+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-tools", "ai-agents", "ai-products", "ai-infrastructure"], "entities": ["Grafana Labs", "Grafana Cloud", "Grafana Assistant", "Grafana Enterprise", "Grafana OSS", "Grafana Play", "YouTube", "Reddit"], "alternates": {"html": "https://wpnews.pro/news/what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs", "markdown": "https://wpnews.pro/news/what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs.md", "text": "https://wpnews.pro/news/what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs.txt", "jsonld": "https://wpnews.pro/news/what-s-next-in-agentic-operations-introducing-ai-week-at-grafana-labs.jsonld"}}