Home
Learning Center
The AI Empowered SRE: Keep Building & Leave the Toil to AI Agents

The AI Empowered SRE: Keep Building & Leave the Toil to AI Agents

Udi Hofesh, Head of Kommunity

3 min read February 12th, 2026

AI SRE Philosophy: If a human operator needs to touch your system during normal operations, you have a bug. AI should be the primary operator for known and recurring operational tasks.

In Site Reliability Engineering (SRE), the core goal is to maximize time spent on long-term engineering projects and minimize time on operational work, which we specifically define as toil. The integration of Artificial Intelligence (AI) and Machine Learning (ML) is the next evolution in achieving this goal. Defining Toil in the Age of AI

Toil remains distinct from purely administrative chores, which fall under overhead (e.g., meetings, HR, goal setting), and valuable grunge work (e.g., cleaning up legacy alerting configurations). For the AI SRE, the definition of toil is sharpened to identify tasks that AI is perfectly suited to eliminate or manage autonomously.

Toil is the kind of work tied to running a production service that typically exhibits the following attributes, making it a prime target for AI automation:

Attribute	AI-Enhanced Description
Manual	Requires hands-on human time, even for executing simple automation scripts. AI Goal: Automate script execution and orchestration, moving to self-driving systems.
Repetitive	Work performed over and over. AI Goal: Identify patterns in recurring issues and automate the entire remediation process (self-healing).
Automatable	A task a machine could do. AI Goal: Implement ML-driven systems to handle predictable, non-essential human judgment tasks, such as intelligent routing, first-level debugging, and automated capacity adjustments.
Tactical	Interrupt-driven and reactive (e.g., handling pager alerts). AI Goal: Predict and prevent issues before they become alerts, or automatically resolve alerts without human intervention.
No enduring value	The service state remains the same after the task. AI Goal: Use the data from toil events to train preventative ML models, ensuring the task is permanently engineered away.
O(n) with service growth	The work scales linearly with service size, traffic, or user count. AI Goal: Leverage AI’s ability to scale sublinearly by managing increasing complexity and load with constant or minimal SRE effort.

Why Minimizing Toil is Critical for AI SRE

The SRE organization maintains the key goal: keep toil below 50% of each SRE’s time. At least 50% must be dedicated to high-level engineering projects. With AI taking on more of the classic toil, SREs are freed up for more strategic work.

The 50% cap and the focus on AI are essential because:

Engineering Enables Hyper-Scale: Utilizing AI to eliminate toil is the “Engineering” in AI SRE. It is the only way the organization can scale to manage the complexity of modern, massive services efficiently.
Keeps the Promise: It reinforces the commitment to SREs that their role is focused on cutting-edge engineering, not operational firefighting.
Toil Expansion: While AI can manage many tasks, unchecked or unaddressed new forms of toil will still consume human time if not proactively engineered out.

What Constitutes AI-Enhanced Engineering Work?

Engineering work is strategic, requires human judgment, and produces permanent, generalized improvements. In the AI SRE context, this work shifts from writing simple automation scripts to designing, training, and maintaining intelligent, self-managing systems.

Category	Description	Examples
Software Engineering	Designing, developing, and refining AI models and the surrounding tooling.	Building ML models for anomaly detection, developing automated remediation pipelines, creating AI-driven infrastructure code.
Systems Engineering	Configuring and documenting systems with a focus on AI integration and governance.	Designing monitoring systems for AI/ML pipelines, defining and tuning Self-Healing systems, consulting on productionization for AI services.
Toil	Repetitive, manual work that AI has not yet, or cannot, fully address.	(As defined above, and continually targeted for AI elimination)
Overhead	Administrative work not tied directly to running a service.	Hiring, HR, team meetings, performance reviews, and training on new AI tools and techniques.

Is Toil Always Harmful in an AI Context?

No. Small, manageable amounts of toil can still provide a valuable feedback loop. However, excessive toil becomes toxic and a systemic failure of the AI SRE program because it signals that the AI systems are failing to automate or prevent known issues. Too much toil leads to:

Impact on Individual SRE	Impact on Organization
Career Stagnation: Time spent fixing things manually is time not spent building the AI systems that would fix them.	Creates Confusion: Undermines the SRE identity as a cutting-edge engineering organization focused on AI-driven automation.
Low Morale & Burnout: Exceeding an individual’s tolerance for performing tasks a machine should have done.	Slows Progress: Human SREs being bogged down in manual tasks halts the development of the next generation of AI-driven features.
	Sets Precedent: Encourages product development teams to rely on the human SRE “safety net” rather than engineering for true operability.

Conclusion

The path to Eliminating Toil is now intrinsically linked to AI-Enhanced Engineering. By committing to a consistent, strategic effort to leverage AI in identifying, predicting, and automatically remediating operational work, SREs can move from operational work to pure, high-value engineering.

Invent more intelligent systems, and toil less!

Latest Articles

The AI-Empowered SRE: AI-Driven Service Level Objectives

e are no longer simply moving bytes; we are managing data ingestion, feature engineering, complex model serving, and real-time inference. To run a service of this magnitude, Site Reliability Engineers (SREs) must move beyond basic uptime metrics and adopt a rigorous, mathematical framework for defining quality.

The AI-Empowered Site Reliability Engineer: Automating the Balance of Risk and Velocity

Extreme reliability comes at a non-linear cost: maximizing stability limits how fast new features can be developed, dramatically increases the operational cost, and reduces the features a team can afford to offer.

What Is AI SRE?

What is AI SRE? How enterprises handle 3x the K8s infrastructure with the same SRE headcount. Autonomous agents eliminate bottlenecks.