What happened
Snowflake held the inaugural CTO Circle event during Snowflake Summit 2026 in San Francisco, bringing together more than 350 CTOs to discuss how to build AI-native engineering organizations. Leaders from financial services, telecommunications, retail, and technology shared insights on three core themes: using AI in production, balancing velocity and risk, and designing engineering teams for AI.
Why it matters
Most organizations start with coding assistants, which improve productivity marginally but leave underlying systems unchanged. The discussion revealed that real competitive advantage comes from treating developer productivity as a product problem, documenting proven AI workflows across teams, and embedding AI into core business operations—not isolating it in dedicated teams. Snowflake's own approach increased its internal developer Net Promoter Score by more than 30 points in 18 months, reaching a 4:1 ratio of satisfied to dissatisfied developers.
What to watch
Leaders emphasized that moving fast requires both strong governance and reliable infrastructure. Netflix's success with automated root cause analysis depended on years of data architecture work before AI agents were introduced. Hex's shift from a centralized AI team to distributed ownership across product teams illustrates the "burning the boats" commitment required—AI must become foundational to how the entire business operates, not a separate function.
Summaries like this, in your inbox every morning.
The CTO Circle conversation reveals a fundamental shift in how engineering leaders think about AI adoption. Early AI initiatives often treated AI as an add-on—deploying a coding assistant and expecting productivity gains. But the speakers at the Summit made clear that this approach leaves organizations largely unchanged. Instead, the most successful companies are redesigning their entire engineering systems around AI as a foundational capability.
Snowflake's experience illustrates this shift concretely. Rather than assuming leadership knew what engineers needed, the company applied product management discipline to the problem: interview users (developers), identify friction, measure outcomes, and iterate. The result was not just faster coding but a transformation in how teams validate ideas—because AI reduces the effort to turn ideas into working software, teams can now prototype, test, and learn by building rather than debating through presentations. This changes the nature of software development itself.
The tension between velocity and reliability also shaped the discussion. As Jeremy Burton noted, AI introduces new operational challenges: agents generate more telemetry, interact with more systems, and make decisions using distributed information. The solution is not to slow down but to invest in architectures that make both speed and reliability possible. Netflix's years of data foundation work before deploying AI agents exemplifies this: rich, accessible context—semantics, ontologies, business rules, standardized APIs—is what allows AI to reason effectively about production environments. Without it, organizations remain stuck with disconnected systems that even AI cannot reason about clearly.
Finally, the discussion highlighted an organizational imperative: companies that compartmentalize AI in dedicated teams create bottlenecks; those that distribute AI ownership across product teams and require full commitment move faster. Hex's pivot from a centralized AI organization to distributed responsibility across every product team captures this shift. The implication is stark—partial AI adoption eventually becomes untenable; organizations must "burn the boats" and restructure how they work.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
