For much of my career, great tax professionals shared deep expertise and a genuine human connection. They navigated complexity, interpreted ambiguity, and saw around corners that others missed.
And they earned the kind of trust that made clients call them first when something changed, because they understood the business, risk tolerance, and goals. That combination was genuinely hard to replicate. It still is.
AI has changed part of that equation. Expertise still matters, but tasks that used to take hours now take minutes, shifting the focus of work from finding answers to interpreting them.
What I’ve watched unfold inside one of the world’s most knowledge-intensive professions points to something every industry is about to face. The advantage was always within the judgment and the relationships behind it. And even as AI becomes more capable, responsibility and differentiation ultimately lie in that judgment.
That shift is already changing how work gets done.
A different kind of readiness
AI agents are now embedded in workflows, handling real tasks, and operating as thought partners. That changes what organizations need from their people. When technology reaches that level of integration, it can’t be confined to a small group of specialists.
The World Economic Forum estimates that nearly 40% of today’s skills could be transformed or outdated by 2030. Leaders can’t afford to leave most of their workforce on the sidelines, so adaptability itself has become a core business capability.
Close the gap
Most companies aren’t there yet. A global IBM study found that a significant barrier to AI success is workforce capability, which explains why some companies move quickly while others stall.
For years, organizations leaned on a core group of enthusiasts willing to experiment while everyone else watched. That made sense when the technology was still emerging. But when capability concentrates in a few people, the company can’t scale.
It’s wise to invest in broad-based upskilling, to create room for curiosity and experimentation at every level, and to build cultures where evolving how you work is a professional strength, whether you’re the analyst two years out of school or the partner who’s been practicing for thirty.
Amplify human judgment
When organizations get this right, AI strengthens human judgment. Judgment is the ability to weigh what the AI surfaced against what it couldn’t know, whether that’s the client’s risk tolerance, the regulatory uncertainty on the horizon, or the relationship context that shapes how you help a client move forward with confidence.
AI creates space for the things that have always mattered most, including interpretation, counsel, and relationships. In a world where information is abundant and cheap, judgment becomes more valuable and more visible than ever before.
The future of AI is about what becomes possible when access is widespread, silos come down, and every employee can actually put it to work. Those that recognize this—and act on it—will move faster because their people adapt faster.
From experimentation to empowerment
We’re past the experimentation phase. AI is here to stay, and the stakes are real. The question has shifted from “What can AI do?” to “What do our people need to do differently because AI exists?”
That requires leaders to rethink how they train, communicate, and measure success. It means putting people at the center of AI strategy. Empowerment happens when learning is part of the flow of work, when judgment is shared rather than centralized, and when leaders make clear that progress is the goal.
The real story
After watching AI reshape the profession where I’ve spent my career, I’ve come to believe that strategic companies will use AI to free their people up for the work that has always been hardest to scale, including judgment calls, difficult conversations, and relationships built on trust over time.
That’s not wishful thinking. It’s what the evidence points to—if leaders are willing to act on it.
Rema Serafi is vice chair of tax at KPMG LLP.