Agent Mode now runs on all popular models, and here's what that actually means.

Next week we’re going to be at Google Next at the Mandalay Bay in Las Vegas, and we’re debuting some incredible work from the groundcover engineering team that brings AI observability to the forefront of agentic workloads and builds on groundcover Agent Mode. 

groundcover, the bring-your-own-cloud-driven (BYOC) observability platform uses an innovative architecture to fill the gaps in legacy SaaS observability, and Agent Mode is the most significant part of the ecosystem to see its benefits. Designed to run in our customers’ own cloud environments and using the customer’s large language model (LLM) services, Agent Mode enables teams to troubleshoot faster, and without exporting sensitive telemetry data out of their environment. That limits security and compliance concerns for groundcover customers, but it also acts as a critical cost control. 

That innovation will be ready to deploy to all clouds starting next week, at the opening of Google Next in Las Vegas, NV. 

Why “inside your cloud” matters

groundcover customers already pay for their own cloud bills, and if we were a SaaS platform they'd have to pay for groundcover's cloud expenses too. By comparison, with BYOC observability, they don't have to. The commercial consequence of BYOC and groundcover deploying to our customers’ clouds is that we can afford to charge customers based on node instead of by volume of data processed. We pass the savings of our innovative deployment on to the companies that need it most, and so they can afford to observe and store all of their observability data. There’s no sampling unless it’s the right choice for your business.

By now, the tech world understands that agentic workflows are only as good as the data (often distilled into ‘context’) the agents have access to. If you’re forced to sample, you’re missing critical data for your context. That’s a tradeoff groundcover customers don’t have to make. Agents enjoy full context, and with agent mode the data has never been easier to search.  

How it works: gcQL (groundcover Query Language)

At the core of Agent Mode is gcQL (groundcover Query Language), a purpose-built query language. It enables the agent to construct precise, complex queries across logs, traces, metrics and events. gcQL significantly improves the accuracy and reliability of AI-driven investigations which is a meaningful technical distinction from approaches that simply layer a chatbot on top of existing observability data. 

As of Google Next (April 22, 2026), Agent Mode will debut within Google Cloud. groundcover already supports Agent Mode in AWS, and across all clouds, Agent Mode conveniently consumes tokens via integrations with Bedrock, OpenAI, and Anthropic. Customers have the ability to define token usage limits, mitigating the risk of overconsumption.

Agent Mode is designed to act as a natural extension of the observability workflow, enabling engineers to ask questions, generate dashboards, create monitors and run queries from a single interface. Its context-aware capabilities allow teams to analyze complex, distributed systems more efficiently across microservices and dynamic cloud environments.

What's new at Next 

In the press release issued this morning, we highlighted that at Google Next, groundcover will also preview new enhancements to Agent Mode including:

• Intelligent, environment-aware prompt recommendations that surface the most relevant investigation starting points — my take on this is that users stand to benefit immensely from the helpful guide posts the platform provides as starting points.
• Trigger-based agentic background workflows that continuously monitor environment signals and autonomously initiate investigations, alerts and remediation actions — my view is that this is exactly what dev/eng teams need to build repeatable and scalable systems instead of just sorting through breadcrumbs every time there’s an issue.
• Cursor integration for AI-assisted code remediation, enabling engineers to move directly from a production issue detected in groundcover to an automated pull request in their repository — this is potentially the most surprisingly powerful shift in how users accelerate their work from within groundcover and looks like a productivity boost out of the box.

In addition to everything mentioned above, there’s an exciting update we’re teasing here for new additions to the core groundcover platform. We're also releasing new observability features specific to AI workloads. We can’t give away too many details. But you should come see it in person. 

Join groundcover at Google Cloud Next 2026

groundcover invites attendees to visit Booth 5301 at the Mandalay Bay Convention Center in Las Vegas to see Agent Mode in action on Google Cloud Platform. 

To learn more or schedule a meeting, visit: Google Cloud Next 2026.