AI SRE roundtable: The creation of a new category

It’s no surprise following the explosion of AI-driven code generation tools, that we are now seeing a strong AI-driven response from operations and security. Almost by definition, a core function of SREs and security professionals alike is keeping up with the pace of change; of being quick to respond in order to maintain the reliability and safety of online systems.

This brings us to the topic at hand: the rise of the AI SRE. There are now three dozen software vendors who either offer an AI SRE agent or are a full-blown, proclaimed AI SRE platform. It seems clear that, while the term certainly has its misnomers and limitations (as all categories do), it’s become a sticky term in 2025, the way “observability” did in 2017.

To help separate the wheat from the chaff, I gathered a roundtable of the foremost thought leaders in operations. The panelists include Kolton Andrus, CEO and Founder of Gremlin; Tucker Callaway, CEO of Mezmo; and Shmuel Kliger, Founder of Causely.

There’s unanimous agreement that AI can effectively enhance operational efficiency. Shmuel Kliger points out that it’s already been doing so for decades, and that “AI SRE” is just the newest term for the latest trends in AI-driven operations.

Kolton Andrus was quick to emphasize that, given the fact that AI can’t be held accountable for high-level strategic decisions, humans will necessarily remain in the loop for the foreseeable future. And while the ability to better anticipate failures and alleviate some manual toil are positives for the industry, there will be a hangover in the new year for companies that over-relied on the AI hype in 2025.

Tucker Callaway stressed that while AI can analyze data faster than traditional methods, costs start to balloon and accuracy begins to fall at enterprise scale. The best way to address this, according to Callaway, is to improve context engineering by preprocessing the data before it’s consumed by agents to maximize efficiency.

There’s an obvious call-and-response here, where new AI solutions are built to keep up with the pace of others, and navigating what is actually useful will be a big challenge for most organizations in 2026. There are also skill gaps, as SREs will need to continuously update their skills to leverage AI effectively.

Andrus suggested that organizations invest in training and development programs to equip their teams with the necessary expertise to work alongside AI tools. Callaway added that collaboration between development and operations teams becomes even more vital in an AI landscape, as both sides need to understand how to integrate AI solutions seamlessly into existing workflows.

Looking ahead, as AI SRE tools continue to mature, they will become more intuitive and user-friendly, enabling broader adoption across various teams and sectors. And it seems inevitable that more and more pieces of the SRE function will become automated in the coming years. While the benefits may be considerable, a measured approach is essential to navigate the challenges posed by these technological advancements.

By fostering a culture of continuous learning, collaboration, and ethical conduct, organizations can harness the full potential of AI SRE in their journey towards operational excellence.

The post AI SRE roundtable: The creation of a new category appeared first on The New Stack.