China Widener, Vice Chair, US Tech, Media & Telecommunications leader at Deloitte – Interview Series

China Widener is Deloitte’s vice chair and US Technology, Media & Telecommunications (TMT) industry leader. She also serves on Deloitte’s US Board of Directors.

She offers a distinctive perspective on the future of the TMT industry, with a particular interest in the evolution of Agentic AI—intelligent systems capable of autonomous decision-making—and their transformative impact across the enterprise. She is passionate about helping clients unlock the full potential of Agentic AI to accelerate innovation, improve operational efficiency, and create new sources of competitive advantage, all while maintaining a steadfast commitment to trustworthy AI.

China has co-authored articles on AI, technology frameworks for enterprises, delivery model analysis, and closing the talent gap. She is a highly sought-after speaker and facilitator and her technology experience and career progression has been featured on several podcasts and publications.

Deloitte is one of the world’s largest professional services firms, providing audit, consulting, tax, risk advisory, and financial advisory services to organizations across virtually every industry. With a global network spanning more than 150 countries, Deloitte works with multinational corporations, governments, and emerging businesses to navigate complex challenges, drive digital transformation, and improve operational performance. The firm is particularly known for its expertise in areas such as technology consulting, data analytics, cybersecurity, and regulatory compliance, helping clients adapt to rapidly changing markets while maintaining strong governance and long-term growth strategies.

Deloitte’s research shows most organizations are still stuck in pilot mode. What is the biggest misconception leaders have about what it takes to scale AI across the enterprise?

Organizations often try to “bolt AI onto” existing processes, when scaling actually requires re-architecting how work gets done. The biggest misconception that I see is that scaling AI is primarily a technology challenge. In reality, the technology is often the easiest part. What leaders tend to underestimate is the degree of operating model change required—from workflows and decision rights to talent, incentives, and governance. 

The shift from isolated use cases to enterprise-level orchestration can be unlocked when people, processes, and technology evolve together. Without that alignment, even the most advanced AI remains stuck in pilot mode.

Why do so many AI initiatives fail to deliver measurable ROI, even when the underlying technology is sound?

Real ROI comes when AI is embedded into end-to-end processes, not treated as isolated experimentation. Most AI initiatives don’t fail because the technology doesn’t work—they fail because the business isn’t set up to capture value from it.

Disconnected use cases, poor workflow integration, and unclear value tracking prevent AI from translating into measurable impact. What’s often missing is orchestration. AI value isn’t just vertical—within a single capability or function—it’s also horizontal, spanning processes, teams, and the enterprise. When organizations don’t connect these layers, they end up with pockets of progress that never scale.

Value comes from aligning technology with workflows and business outcomes so AI can operate across the enterprise, not just within silos. That’s when you move from experimentation to true transformation and start realizing sustained, cross-functional value.

At what point should a company stop experimenting and commit to enterprise-wide AI transformation, and what signals indicate they are ready?

Organizations should move beyond experimentation when they’re ready to embed AI into end-to-end processes. That’s a shift from treating AI as a collection of bespoke pilots to viewing it as a strategic transformation anchored in a clear roadmap for where value exists and how to capture it.

That means aligning stakeholders around shared priorities, defining where AI can drive the greatest impact across functions, and putting the right governance and operating model in place to execute against it. Without that clarity, organizations remain stuck with pilots that are difficult to scale or replicate.

In short, success will come down to moving from fragmented experimentation to a coordinated, enterprise-wide strategy—where value is clearly defined, prioritized, and systematically realized.

How should executives rethink ROI in the context of AI, especially when benefits extend beyond cost savings into productivity, workforce redesign, and long-term strategic advantage?

It’s about thinking of AI ROI as more than just cost savings. When you move into a broader value equation that includes productivity gains, workforce transformation, and new avenues for growth, that’s when you unlock AI’s true value. While efficiency matters, the larger impact often comes from redesigning roles, accelerating decision-making, and enabling entirely new business models. The challenge is that these benefits don’t always show up in traditional financial metrics or short-term P&L cycles. The key is adopting more holistic measurement approaches that combine financial, operational, and workforce outcomes. Ultimately, the goal is to evaluate AI not just as a cost lever, but as a driver of long-term competitive advantage.

One of the challenges highlighted is the gap between visibility and action. Why do organizations struggle to operationalize insights generated by AI?

The gap between insight and action is an execution problem, not an analytics one. Organizations can sometimes generate insights, but unclear decision rights, misaligned workflows, and lack of accountability prevent action. Governance is key. Closing that gap requires embedding AI and understanding the impact to workflows, practices, policies and execution paths.

Enterprise AI Navigator emphasizes the “agentification” of tasks. How should leaders decide which processes are best suited for AI agents versus human-driven workflows?

Different processes require different levels of automation; effective transformation is about making the right choices. The key question isn’t whether to deploy AI agents—it’s where they create the most value.

That value can be evaluated in two primary ways: organizational “fit” and financial impact. For some organizations, alignment with existing workflows, culture, and ways of working will be the priority, making “fit” the best starting point. For others, the focus will be on measurable financial returns, where understanding the range and scale of value creation becomes the primary driver. The ability to assess both dimensions is what enables more informed, strategic decision-making.

It’s important to note that not every process should be agentified. Some require human judgment, trust, or creativity. The goal is a hybrid model, where humans and agents are intentionally designed to complement each other.

Many organizations are investing heavily in AI tools but not redesigning workflows. How critical is organizational restructuring to achieving real AI impact?

Organizational restructuring isn’t optional; it’s central to achieving AI impact. Companies that invest in tools without redesigning workflows typically see incremental gains at best. Real value comes when organizations rethink roles and responsibilities, team structures and how decisions are made. AI doesn’t just change tasks: it changes how work flows across the enterprise. Without structural alignment, true transformation is limited.

Governance is lagging behind adoption, particularly with agentic AI. What risks are companies underestimating as they scale more autonomous systems?

Our research shows that only 21% of organizations have mature governance in place for autonomous agents. Without a strong, end-to-end governance program many companies hesitate to deploy these tools. That hesitation often leads to the rise of “shadow AI,” which introduces significant unmanaged risk. 

At the same time, organizations frequently underestimate the role employees can play in governance. Policies and reporting structures are critical, but they’re not enough on their own. Employees generally want to use AI responsibly—and when they’re equipped with clear guidance, they become a powerful line of defense.

That means organizations should focus on educating their workforce: what’s safe, what’s risky, and how to make good judgment calls in real time. For example, pausing to ask: should I include sensitive financial data here? That kind of day-to-day decision-making is where governance truly comes to life. Employees, when informed and empowered, actively reinforce the organization’s risk posture.

Finally, ongoing monitoring and periodic testing remain underutilized but are essential. As we move further into the era of agentic AI, continuous oversight needs to become table stakes for any organization deploying these capabilities at scale.

Deloitte suggests Enterprise AI Navigator can significantly reduce strategy and design time. What specifically changes in how organizations approach AI decision-making when using a system like this?

Enterprise AI Navigator connects financial, workflow, and workforce insights into a unified transformation roadmap. What changes is the shift from intuition-led decision-making to data-informed transformation that reflects a disciplined approach to use across the stakeholder landscape. With Enterprise AI Navigator, Deloitte can help model scenarios before investing, make AI decisions tied directly to financial and workflow impact, and move from isolated pilots to a cohesive, enterprise roadmap

It effectively compresses strategy and design cycles by giving leaders visibility into what will work—before they scale it.

Looking ahead 1 to 2 years, what will separate the companies that successfully turn AI into a competitive advantage from those that remain stuck in experimentation?

The divide won’t be about who adopted AI; it will be about who transformed because of it. Those leaders will treat AI as a business transformation lever, not just a toolset. That means redesigning workflows and operating models end-to-end, while measuring value holistically across financial, workforce, and growth outcomes.

Crucially, they’ll recognize that productivity alone isn’t the finish line. Many organizations are already seeing incremental efficiency gains, but competitive advantage comes from using AI to unlock new revenue streams, reshape offerings, and drive enterprise-wide growth, not just do the same work faster.

Those that lag will continue to focus on tools over outcomes, chasing isolated use cases without connecting them across the business. They’ll also tend to underinvest in change management, governance, and orchestration—making it difficult to scale impact.

In short, the winners will be the ones that move from AI-enabled to AI-powered enterprises. These enterprises will embed AI is into how the business runs, how it grows, and how it competes, not just how it experiments.

Thank you for the great interview, readers who wish to learn more should visit Deloitte.