Key Takeaways

  • Resolve AI has exited stealth with a platform designed to function as an autonomous Site Reliability Engineer (SRE).
  • The startup secured $35 million in seed and Series A funding led by Greylock, with participation from notable angel investors.
  • The technology distinguishes itself by focusing on autonomous remediation and action rather than just alert summarization or observability.

For most engineering teams, the promise of automation usually hits a hard ceiling: the moment a decision needs to be made. We have endless tools to monitor infrastructure, log errors, and route alerts to PagerDuty, but when the server actually crashes at 3 a.m., a human still has to wake up, log in, and fix it.

Resolve AI, a startup emerging from stealth, is betting that the technology is finally mature enough to remove the human from that loop entirely.

The company has introduced what it calls an autonomous site reliability engineer (SRE)—a system designed not just to flag issues, but to troubleshoot and resolve them without human intervention. To back this vision, the startup has raised $35 million in combined seed and Series A funding, led by Greylock.

It’s a bold pitch in a market saturated with "AIOps" tools that often amount to little more than sophisticated pattern matching. But Resolve AI’s approach—and the pedigree of its founders—suggests they are building something fundamentally different. The company was founded by Spiros Xanthos, a veteran of the observability space who previously founded Omnition (acquired by Splunk) and served as VP of Product Management at Splunk.

Moving Beyond the Copilot Model

The current wave of generative AI has largely focused on the "copilot" model: an assistant that sits beside an engineer, offering code suggestions or summarizing logs. It’s helpful, sure, but it still requires a pilot.

Resolve AI is pushing for an agentic model. The system integrates with existing cloud infrastructure (AWS, Kubernetes) and observability stacks to build a causal graph of the environment. When an incident occurs, it doesn't just ping a Slack channel. It investigates the dependencies, identifies the root cause, and executes a remediation plan.

It’s a small detail, but it tells you a lot about how the AI market is shifting: we are moving from tools that chat with us to tools that work for us.

The platform is designed to handle the tedious, repetitive incidents that cause widespread burnout among SREs—things like restarting stuck pods, rolling back bad deployments, or clearing disk space. By offloading this "toil," as Google’s SRE handbook famously calls it, teams can theoretically focus on architecture rather than firefighting.

The Capital Behind the Code

The $35 million in funding is significant for a debut, signaling strong investor confidence in the "autonomous worker" thesis. Alongside Greylock, the round saw participation from heavy hitters in the developer infrastructure space, including Stanislav Vishnevskiy (co-founder of Discord) and other industry angels.

This level of backing suggests that venture capital is looking for the next layer of infrastructure software—one that doesn't just manage complexity but actively reduces the operational headcount required to maintain it.

Trusting the Machine

Of course, the technology triggers an immediate, pragmatic question: Are engineering leaders actually ready to let an AI write-access their production environment?

It’s one thing to let an LLM write a unit test; it’s another to let it execute a rollback on a live transaction database. That’s where it gets tricky. Trust is the primary bottleneck for autonomous agents in high-stakes environments.

Resolve AI seems aware of this friction. Their system operates with a "human-in-the-loop" mode initially, allowing engineers to approve actions before they are taken. Over time, as the system proves its accuracy, teams can switch specific categories of incidents to fully autonomous mode. This graduated approach is likely the only way to get conservative enterprises on board.

The Technical Underpinnings

Unlike generic chatbots, Resolve AI’s "SRE" is built on a proprietary reasoning engine tailored for IT operations. It understands the specific topology of microservices and the cascading nature of failures.

When an alert fires, the system attempts to replicate the deductive process of a senior engineer. It checks recent deployments, looks for latency spikes in dependent services, and correlates logs across disparate tools.

Still, the challenge remains integration. Every company’s stack is a unique combination of legacy code, multi-cloud configurations, and undocumented workarounds. An autonomous engineer is only as good as the data it can access and the tools it is permitted to touch. If Resolve AI encounters a proprietary internal tool with no API, its autonomy hits a wall.

The Shift in Operational Logic

The emergence of Resolve AI highlights a broader trend in the B2B sector. We are seeing a transition from software that provides visibility to software that provides agency.

For years, the industry mantra was "single pane of glass"—the idea that if you could just see everything in one place, you could fix it faster. But data volume has exploded to the point where visibility alone isn't enough. Seeing a thousand red alerts on a dashboard doesn't help if you only have three engineers to address them.

Resolve AI is betting that the solution isn't a better dashboard, but a synthetic team member who doesn't sleep.

The startup is currently deploying its technology with early design partners. Whether it can deliver on the promise of a truly hands-off production environment remains to be seen, but the capital and talent behind the project indicate that the hunt for the autonomous SRE is officially on.