KPMG careers hinge on human skills as AI takes routine work

The new value in a KPMG career is not faster production, it is better judgment

AI is already eating into the routine work that used to define many audit, tax, and advisory hours. That changes the career equation inside KPMG: the people who stand out will not be the ones who simply generate more output, but the ones who can decide what matters, explain why it matters, and keep clients steady when the answer is not obvious.

The Journal of Accountancy’s May 1 feature captures that shift plainly. It says the five human-centered competencies that remain most valuable are critical thinking, creativity, empathy, strategic vision, and strong interpersonal relationships. For KPMG professionals, that is not abstract theory. It is the difference between a staffer who can surface a variance and a senior who can tell a client whether the variance is noise, a control issue, or a problem that could change the engagement.

Why those five skills matter in real client work

Critical thinking becomes more valuable when AI can summarize a ledger, draft an initial memo, or spot a pattern before a human does. The human job moves to the harder part: testing assumptions, noticing what the model missed, and deciding whether an anomaly is material or merely distracting. In audit, that might mean knowing when a weird journal entry deserves a deeper look instead of a quick tick mark.

Creativity matters too, especially in advisory and tax, where the answer often has to fit a client’s business model, risk appetite, and timing pressure. AI can propose options, but it cannot always recognize which option a nervous controller, a tax director, or a CFO will actually accept. That is where strategic vision and interpersonal skill start to overlap: the best professionals do not just solve for technical correctness, they move the client toward a workable decision.

Empathy is becoming a practical workplace skill, not a soft extra. Tax teams are dealing with clients who are overwhelmed by changing rules, and audit teams are often delivering uncomfortable news under deadline pressure. A professional who can read the room, calm a tense meeting, and explain a risk without escalating panic is still delivering something software cannot.

KPMG’s own numbers show why the pressure is rising

KPMG’s 2026 AI in Finance survey says active AI use across the finance function has more than doubled in two years, and more than three-quarters of organizations are already using AI in finance. The survey was based on 1,013 senior finance leaders across 13 sectors and 20 countries, fielded in March 2026. That means this is no longer a pilot phase or a side project. It is becoming the operating environment.

KPMG also frames the shift around trust, governance, controls, and human oversight, which is exactly where human skill becomes a performance issue. If AI is acting as a “decision engine,” then the people around it need to know when to challenge the output, how to document the reasoning, and how to explain the decision to a client, regulator, or internal reviewer. In a firm like KPMG, that shows up in review quality, engagement risk, and whether a manager is seen as someone who can be trusted with more complex work.

Audit teams will need better judgment, not just better tools

KPMG’s earlier study on financial reporting and audit makes the same point from another angle. In that research covering 1,800 companies in 10 countries, 64% of companies said auditors would have a role in evaluating AI use in financial reporting, nearly 72% were already piloting or using AI in financial reporting, and 57% expected to implement generative AI over the next three years. North America was leading at 39%, ahead of Europe at 32% and Asia Pacific at 29%.

Those figures point to a simple reality: audit teams are moving into a world where they are not only testing controls around AI, but also judging whether the AI itself has become part of the reporting process. That raises the stakes for skepticism, documentation, and the ability to spot when a polished output is hiding a weak assumption. The best auditors will not just accept a model’s answer more quickly; they will know when a model’s confidence should make them more suspicious.

Carl Mayes of the AICPA put the shift in plain language in March, saying the profession is moving from “doing” to “supervising” as bots take over repetitive tasks such as vouching. That matters for career development inside KPMG because promotion is no longer only about volume, turnaround time, or technical completion. It is also about whether a professional can supervise AI-generated work, test its logic, and explain the consequences of getting it wrong.

Tax teams are being pushed toward judgment and communication

KPMG’s Tax AI Accelerator Program, launched on February 3, 2026, shows how quickly the tax function is changing. The program uses KPMG’s Digital Gateway platform, built on Microsoft Azure OpenAI, and is designed to help tax teams build practical AI skills and bring generative AI into day-to-day operations. KPMG said Duke Energy is one of more than a dozen participating companies.

The business case is clear in the firm’s own numbers: more than half of tax departments are already using or exploring generative AI, and 86% of tax leaders believe these tools will help address talent gaps. That does not mean tax work becomes easier in the human sense. It means the routine parts can be compressed, leaving more of the profession’s value tied to interpretation, communication, and risk management.

For a tax professional, that changes what “good” looks like in a client meeting. A strong performer is not just the person who can produce a clean analysis. It is the person who can explain what the analysis means for cash, compliance, and future planning, while keeping the client from making a rushed decision on incomplete information. In a busy season environment, that communication skill can be the difference between an efficient response and a costly misunderstanding.

How this should change performance conversations at KPMG

The human premium in an AI-heavy workplace is not a slogan. It should show up in coaching, staffing, and promotion cycles. If AI is handling more drafting, extraction, and first-pass analysis, managers should be asking whether a professional can do four things well: challenge the machine’s output, turn facts into direction, build confidence with stakeholders, and connect finance, operations, and risk.

That matters for partner-track careers because the job gets broader as the work gets faster. The future top performers will be the ones who can move from analysis to judgment, from judgment to client trust, and from client trust to repeat business. AI can help KPMG teams scale technical throughput, but it cannot replace the human skills that keep clients listening when the answer is messy, the pressure is high, and the stakes are real.

The message for KPMG employees is blunt: technical fluency is now table stakes, but it is no longer the whole game. The people who advance will be the ones who can think independently, read the room, and catch the risks that a model can miss.