KubeCon + CloudNativeCon North America 2025 hosted an exclusive end-user summit for cloud-native leaders and innovators. This gathering provided a unique opportunity to network with industry experts, exchange ideas, and learn about the latest trends in cloud native technology.
This year’s theme, “Engaging with AI,” brought together technology leaders from top enterprises to explore how organizations are adopting and scaling artificial intelligence responsibly and effectively. The afternoon featured fireside chats with executives from End User member companies, followed by focused breakout discussions on vendor selection, scaling AI initiatives, and demonstrating ROI. The conversations revealed both the promise and the complexity of enterprise AI, highlighting the tension between rapid innovation and structured governance. The event underscored a shared theme: AI transformation is much more about organizational agility and culture than it is about technology.
Members of the Artificial Intelligence Technical Community Group attended the event and participated in multiple conversations. This blog highlights the pain points, patterns, and solutions that emerged from the discussions.
Pain points
Despite enthusiasm for AI’s potential, most organizations are struggling to reconcile innovation with operational constraints. The summit identified several recurring pain points, including vendor management, governance, ROI measurement, standardization, and driving adoption. These challenges are often interdependent, slowing execution even when leadership support is strong and the budget is available.
1. Vendor volatility and dependence
AI ecosystems evolve quickly. Enterprises risk losing critical partners when vendors are acquired or shut down. To mitigate this, some teams are experimenting with abstraction layers and developing SDKs on top of third-party SDKs to insulate their systems from vendor instability.
2. Governance and bureaucracy
Heavily regulated industries face a friction point between the speed of innovation and the rigor of compliance. Traditional vendor approval cycles, often lasting eighteen months or longer, cannot keep pace with the rapid development of AI. In addition, measuring new tooling and its subsequent efficiency gains is cumbersome and often imprecise. Leaders emphasized the need for agility within strong governance frameworks.
3. Unclear ROI and measurement metrics
Organizations are measuring the wrong things. Many report adoption rates (e.g., prompt count, Copilot usage) rather than business outcomes (e.g., time-to-market reduction, developer productivity). This disconnect makes it difficult to sustain executive buy-in and justify continued investment. While the focus of management consulting organizations is to calculate micro-tasks as a means to evaluate end-to-end efficiencies and gains, these methodologies are rarely applied in practice today.
4. Human factors and change management
AI deployment often stalls due to human resistance, skill gaps, or a lack of structured learning opportunities and evaluation capabilities to justify investment returns. Both large telecoms and banking teams emphasized that AI transformation mirrors past cultural shifts, such as DevOps and Agile, requiring empathy, communication, and psychological safety as much as technical training.
5. Product lock-in and limited flexibility
Dependence on dominant ecosystems (such as NVIDIA or Microsoft Copilot) creates cost pressures, business continuity risks, and limits diversification. Enterprises want flexibility and in some cases sovereignty, but find that switching vendors introduces integration complexity and revalidation costs.
Patterns
Across the summit’s discussions, several clear patterns emerged in how enterprises are adapting their AI adoption strategies. These patterns reveal a shift from exploratory experimentation toward intentional, scalable frameworks that blend autonomy with centralized oversight.
1. Shift from experimentation to execution
Organizations are maturing from isolated PoCs to institutionalized AI frameworks. Many now have formal councils or governance boards approving AI use cases, creating repeatable processes for scaling successful pilots.
2. Cross-functional collaboration
The most successful AI efforts are built through cross-functional partnerships between technology, product, legal, and risk teams. This collaboration enables faster decision-making without compromising compliance or security. A joint-led approach has proven successful in balancing strategic direction and time-to-value.
3. Human-centric transformation
Enterprises are recognizing that technology adoption only succeeds when humans are empowered, provided the right flexibility, tooling, and time to work with their new tooling, and adopting fail-fast methodologies for experimentation and evaluation. Internal training frameworks integrate change management, coaching, and improvements to developer experience alongside AI tooling.
4. AI as infrastructure, not an add-on
AI is increasingly viewed as a foundational layer, embedded within the horizontal tech portfolio, rather than an isolated innovation project. This normalization of AI across compute, networking, and observability layers signals strategic relevance, with a continued need to incubate and invest in this technology domain and workstreams to remain agile and accountable, while fending off competition in core business domains.
5. Outcome-based scaling
AI initiatives are now being evaluated against tangible performance indicators such as DORA metrics, NPS, and time-to-market acceleration. However, the audience noted that these metrics remain imperfect at providing reliable, industry-wide measures. This shift toward outcome-based scaling aligns AI efforts with business value and trust.
Solutions
Participants shared practical strategies for overcoming barriers and operationalizing AI adoption. These solutions combine process redesign, technology standardization, and cultural reinforcement, showing that sustainable transformation requires both structural and behavioral change.
1. Frameworks for vendor resilience
Develop layered architectures that abstract away vendor dependencies to facilitate business continuity. Encourage redundancy given the future criticality of AI capabilities to the business, while focusing on open standards to minimize disruption if vendors pivot, raise costs, or go out of business.
2. AI governance and use-case approval
Create governance councils that evaluate AI use cases early, with clear criteria for security, compliance, and business relevance. While maintaining and fostering innovation within the business. Decoupling experimentation from production deployment to preserve agility while maintaining oversight and control remains essential.
3. Human enablement and change management
Invest in structured learning programs to provide hands-on experimentation and cross-team learning. Celebrate internal success stories to drive adoption through peer influence and recognition.
4. Defining and measuring ROI
Shift measurement frameworks from usage metrics to business impact metrics. Track developer velocity, feature delivery time, software quality, security metrics, and customer satisfaction as indicators of AI’s real value. Use baseline experiments to quantify improvements.
5. Balancing agility with compliance
Empower business units to experiment within shared security and governance guardrails. Encourage smaller, iterative PoCs to learn quickly, while maintaining centralized risk oversight for large-scale deployments. With regular re-evaluation based upon the cadence of change within the technology domain.
Conclusion
The summit made clear that enterprise AI adoption is entering a new phase: one defined by institutional integration rather than isolated experimentation. The discussions emphasized that the true challenge is not technical feasibility but organizational readiness, aligning people, processes, and technology to achieve measurable business outcomes.
The leading organizations view AI not as a feature but as a capability that requires cultural transformation, governance innovation, and strategic patience. As the summit participants demonstrated, success will belong to those who build resilient, human-centered systems that can evolve in tandem with the technology itself.