Practical AI in Platform Engineering: Lessons From Port's Latest Meetup

Why It Matters

The findings highlight a gap between AI hype and operational maturity, signaling that most engineering orgs need robust evaluation and governance frameworks before scaling agentic workflows.

Key Takeaways

• 67% of teams run a few agents in SDLC, not at scale
• Code review (94%) and generation (89%) are most shipped AI features
• Only 10% have automated CI evaluations for agents
• Governance and guardrails concern 61% of engineering leaders

Pulse Analysis

The rise of generative AI has turned platform engineering into a testing ground for autonomous agents that can write, review, and even deploy code. Companies such as Port, monday.com, and HiBob are experimenting with these tools, but the broader ecosystem remains in a transitional phase. A recent meetup in Tel Aviv gathered senior engineering leaders and distilled their experiences into a concise survey, offering a rare snapshot of how AI‑driven workflows are being integrated into real‑world development pipelines.

The survey paints a nuanced picture of maturity. While two‑thirds of respondents report using a handful of agents in the software development lifecycle, only 17% claim fully autonomous AI‑SDLC capabilities. Code review dominates adoption at 94%, followed closely by code generation at 89%; security and infrastructure automation lag behind, hovering below 40%. Evaluation practices are equally uneven—28% rely on informal “eyeball” checks, and a mere 10% have automated continuous‑integration tests for their agents. Human‑in‑the‑loop controls remain centered on pull‑request merges, reflecting a cautious approach to production‑level autonomy.

The consensus is clear: governance, guardrails, and access control are the primary bottlenecks, cited by 61% of participants. Without standardized policies and robust monitoring, scaling agentic workflows risks technical debt and unpredictable behavior. Organizations that invest early in structured evaluation frameworks and cross‑team governance are likely to reap the productivity gains promised by AI while mitigating risk. As the industry moves beyond the coding‑assist phase toward full‑stack automation, the lessons from this meetup serve as a roadmap for engineering leaders seeking to balance innovation with operational control.