Kubecon Survey: How Platform Teams Are Adopting Ai And Idps

Platform engineering continues to beryllium cardinal to nan occurrence of package organizations. During KubeCon successful November, we surveyed 219 level engineers connected what’s apical of mind. The study responses springiness a glimpse into wherever level teams beryllium successful nan organization, who they service and what they are prioritizing next.

Platform’s Primary Customer Is Still Developers

What Steve Ballmer said 25 years agone still holds true: “developers, developers, developers, developers, developers…developers.” If there’s 1 connection that comes done large and clear successful study responses it’s this: Platform engineering is still, supra all, a developer merchandise team.

• Developers predominate arsenic nan superior customer: 44.1% of respondents selected developers arsenic their main “customer.”
• SRE / Infra teams travel adjacent astatine 20.2%, reinforcing that galore platform teams besides transportation a reliability and operational enablement mission.
• The customer guidelines is widening: Data teams (13%) and information teams (10.4%) are besides meaningful stakeholders.

Drilling into nan 2nd 2 bullets: Even erstwhile nan level serves data, information aliases infrastructure directly, nan shared result is still a faster, safer way to accumulation for developers. This comes successful nan shape of automating information tasks, database-focused operations and driving efficiencies passim nan broader package transportation life rhythm (SDLC).

The operation of “customers” hints astatine a displacement successful really level teams deliberation astir their raison d’être. The level isn’t conscionable a paved roadworthy for shipping code; it has go a shared soul product, 1 that must meet aggregate disciplines without losing show of developer outcomes. As those customer groups grow, truthful does nan request for clear ownership, discoverability, guardrails and standards crossed services, infra, information assets and policies. All successful nan sanction of making judge nan developer tin much efficaciously do their superior occupation (ship awesome code).

AI Isn’t Quite Everywhere [Yet]

The study shows AI is still nascent passim galore organizations. Though, AI has crossed nan statement from “interesting someday” to “we request this now.” In fact, astir each respondents highlighted AI exploration arsenic a apical 2026 priority.

What are they trying first?

• AI codification assistants lead: 30.1% of level teams are exploring really to usage AI coding assistants to amended alteration developers successful 2026.
• Next wave: AI codification reviews (17.8%) and automated summons solution (13.4%).
• Reliability automation is rising: Self-healing incidents (12.2%), trial procreation (10.2%) and auto-fixing vulnerabilities (9.6%) show a hefty appetite for AI that reduces toil.

This shape is grounds that teams are testing AI wherever it astir straight elevates developer throughput and cuts operational drag. Organizations are reasoning astir a “better together” communicative for humans positive AI to create compelling agentic engineering usage cases. Over time, these experiments will push level teams to standardize really AI devices are governed, integrated and measured.

IDPs Are Becoming nan Operating Layer

IDPs are nary longer an emerging concept, but take is uneven.

• Nearly 60% of respondents study utilizing an IDP today.
• There is fragmentation successful how IDPs are used. When asked really developer outcomes are measured, 43% of respondents stated they person nary formalized strategy to cod feedback from engineering to place clash points. Surveys are nan astir communal measurement to measurement developer success, followed by circumstantial metrics and interviews.

What is absorbing is that platform teams specify their ain occurrence metrics based connected developeroutcomes. Reduction of tickets created for DevOps/infra (16.5%) and Deployment Frequency (13.6%) are nan apical 2 KPIs level teams clasp themselves accountable to. Having an IDP is important to streamline this measurement and trim instrumentality sprawl; making nan level teams’ life easier which manifests for amended developer outcomes.

Conclusion

IDPs stay a captious instrumentality for modern engineering orgs: They streamline workflows and bring bid to a chaotic SDLC. But successful nan AI era, transportation is moreover much chaotic and consequential. Without a level furniture that accounts for and enables AI-driven work, teams consequence invisible bottlenecks, brittle automation, protector AI exposure, and slower delivery.

The IDP is evolving to meet that moment. This improvement comes successful nan shape of agentic engineering platforms (AEPs) that widen nan portal into nan AI conception of nan SDLC, supporting not conscionable codification creation, but self-healing incidents, vulnerability remediation, agentic effect measurement and standards enforcement astatine scale.

Group Created pinch Sketch.