Cloud Strategy for AI: Insights from EY’s Global Vice Chair

As AI workloads grow in scale and complexity, the real challenge for enterprises is no longer adoption – it is building environments where AI can operate efficiently, scale responsibly, and deliver tangible business impact. From cloud architecture and workload distribution to governance models and workforce readiness, today’s technology leaders must navigate a delicate balance between innovation, control, and long-term strategic impact.

To explore this topic further, we connected with Mary Elizabeth Porray ( EY’s Global Vice Chair, Client Technology & COO, Growth and Innovation) to understand what it truly takes to build an AI-ready enterprise. In this conversation, she discusses the cultural and skills shifts required for AI adoption, the strategic decisions organizations must make, and the importance of designing cloud environments that can evolve alongside AI ambitions.

Packed with forward-looking insights, this interview is a must-read for leaders aiming to build resilient, AI-ready organizations.

Enterprise AI and cloud strategy

You often advocate a “cloud-fit” approach to enterprise AI. How do you define the characteristics of a cloud environment that is truly AI-ready. And what are the common missteps enterprises make when attempting to scale AI workloads in the cloud?

Porray: A cloud environment is truly AI-ready when it goes beyond infrastructure and actively enables intelligent work. That means seamless data integration across systems, native access to AI services and platforms that reduce operational friction. So teams can move from maintaining technology to using AI for decision-making and innovation.

An AI-ready cloud isn’t just scalable. It’s designed to embed intelligence into everyday enterprise workflows and elevate IT from execution to strategy.

A common misstep enterprises make when attempting to scale AI workloads is assuming that the size of their cloud footprint is a measure of success. Instead, enterprises must understand the value of focusing on cloud maturity that provides elasticity, availability, and speed, not just scale.

This includes redesigning cloud, AI, and data strategies into a single operating model and engineering AI-driven cloud environments with trust and compliance while avoiding unnecessary operational complexity.

Many organizations struggle with deciding where AI workloads should live. What factors, technical, operational, and cultural, should guide these decisions, and how do you balance flexibility with governance?

Porray: The organizations that will lead are those that continuously measure what’s working, refine workload placement based on evidence rather than trends, and build adaptable cloud foundations that can evolve as quickly as their AI ambitions. This combination of intentional architecture, AI-enabled operations, and disciplined measurement is what will define true cloud maturity. AI adoption requires companies to rethink workload placement and be increasingly intentional about what belongs in the cloud. Within cloud environments, legacy architectures can restrict the speed and scalability of AI. Making workload placement a strategic choice rather than a cost consideration.

To balance flexibility with governance, leaders need to be transparent and make oversight clear. It is key for organizations to treat governance as mandatory and a sustainable growth accelerator.

Can you share examples where organizations successfully adapted their infrastructure to keep pace with rapidly expanding AI use cases?

Porray: EY recently prioritized adaptable infrastructure in its own operations through its adoption of Microsoft Azure AI Document Intelligence on its Global Tax Platform.

While the tool includes some prebuilt models, the complexity of processing and pulling data from tax forms can be uniquely challenging. This presented a barrier to EY taking full advantage of Microsoft’s AI Document Intelligence capabilities.

To solve this, EY built and trained its own machine-learning model. With generative AI, EY could annotate tax forms and produce synthetic data to accurately train the model and ultimately integrate Microsoft’s AI use cases seamlessly into workflows.

EY used this approach to reduce manual interventions on the Azure Document Intelligence tool by 80% and process an estimated 16,000 federal and state tax forms.

Human-Centered Transformation

You’ve highlighted that human roles are being redefined, not diminished, in the AI era. How can leaders ensure their workforce is prepared for this transformation while maintaining accountability and innovation?

Porray: As AI becomes more embedded into society, and related threats and incidents grow, trust will become a business imperative. Leaders must prepare their workforce for this shift by recognizing that the future of client and stakeholder relationships will depend on trust and interpretation. And not just data accuracy.

Progress in AI is not just about the tools, but about how leaders help their teams grow with them. We often see concerns from leaders around skills gaps, technical confidence barriers, and the fear of making a mistake that holds employees back. Misaligned messaging at the leadership level could lead to top-down and bottom-down tension, where employees misunderstand expectations and experiment without guardrails (shadow AI). 

Leaders can ease this friction by acknowledging adoption friction directly and addressing people-level concerns openly. Organizations must then establish clear expectations for how AI is used. They should create safe spaces for experimentation and questions, while investing in training that ties AI directly to real job functions, so employees understand its impact. Bringing diverse perspectives into each stage of the process helps drive stronger solutions and smoother transitions.

From experience, I recommend teams create hands-on opportunities to build comfort and curiosity around AI. This can include broad access to copilots, in-house large language models, and “agent builder” tools.

As enterprises scale AI, what human skills or cultural shifts become critical to ensure humans remain central to decision-making and strategy? Rather than being sidelined by technology?

Porray: As enterprises scale AI, the human touch becomes a competitive edge. Judgment, empathy, relationships, and understanding of nuance will be key skills. The most effective workplaces will pair AI automation with a human touch, building ethical, responsible, and people-centered AI uses.

AI can reclaim hours each week that used to go toward non-value-added tasks, freeing people to focus on creativity and relationships.

The goal is not to do less, but to have more energy for the work that truly matters. With AI handling routine work, expectations for people are changing. Knowledge alone will not set anyone apart; instead, emotional intelligence, creativity, and communication are becoming the new differentiators.

Forward-looking Perspectives

What emerging AI workloads or patterns do you believe will shape enterprise infrastructure over the next five years? And how should organizations future-proof their cloud foundations today?

Porray: In the next five years, AI workloads will become even more demanding. This shift will simultaneously heighten regulatory expectations, prompting organizations to confront the reality that AI workloads can’t all reside in a single location.

To future-proof their cloud foundations, organizations must prioritize a “cloud-fit” approach over a traditional “cloud-first” strategy to better manage increasing AI workload demands. This means prioritizing architecture design optimized for elasticity and speed.  

AI will also change how enterprises operate and manage their cloud environments on a daily basis.

Finally, what one guiding principle would you like to share with future enterprise leaders?

Porray: Based on my own experiences evolving from a hands-on “fixer” to enabling coach, I believe the AI era calls for leaders who equip teams to adapt, learn, and make technology work for them. Combining human insight with advanced tools will determine which organizations use AI to become more thoughtful, creative, and connected, ultimately thriving in the next wave of innovation. 

Key takeaway

As Mary Elizabeth Porray explains, scaling AI successfully requires enterprises to adopt a ‘cloud-fit’ approach over a traditional ‘cloud-first’ strategy – one in which architecture, data, and operations are intentionally designed to support intelligent workloads.

At the same time, enterprises must treat governance and trust as enablers of innovation rather than barriers, ensuring AI systems operate transparently and responsibly. Equally important is the human dimension: upskilling teams and giving them the confidence and freedom to experiment with AI while maintaining clear guardrails.

Ultimately, organizations that bring these elements together – adaptive infrastructure, disciplined governance, and empowered people – will be best positioned to turn AI ambition into lasting business value.