How EY's Simon Brown Is Rewiring HR for AI That Doubles Every Six Months

Agentic AI capabilities are doubling in potency every six months, and HR leaders who treat this as just another tech upgrade will soon find their organizations left behind. That’s the stark warning from Simon Brown, EY’s global learning and development leader, who has spent the last two years building a 400,000-employee readiness program to meet the coming wave of autonomous AI agents.

In a recent interview, Brown laid out a pragmatic blueprint that goes far beyond model benchmarks and feature lists. The real barrier, he insists, isn’t the technology—it’s culture, governance, and the way HR designs careers and incentives. “Is the culture there to support this?” Brown asks. “Most organizations are feeling their way through… There’s a lot of ambiguity.”

His message lands at a pivotal moment. Microsoft has confirmed that Copilot now has access to GPT-5 in some configurations, a leap that dramatically improves agent reasoning and planning. Meanwhile, the World Economic Forum projects 170 million new roles by 2030, against 92 million displaced—a net gain but enormous churn. For HR departments still anchored to annual performance cycles and static job descriptions, the clock is ticking.

The Velocity of Change: Why Six Months Matters

Brown’s observation that “AI capabilities double every six months” isn’t hyperbole—it’s an operational planning heuristic backed by industry trends. Satya Nadella has described a “doubling on the order of months,” and independent benchmarks show task-time horizons halving every 4–7 months for certain agentic tests. In practice, this means a tool you tested 90 days ago may already be obsolete.

That acceleration caught many executives off guard. Procurement cycles often lag 6–12 months behind the public release of new models, creating a dangerous knowledge gap. Brown warns that leaders must stop thinking of AI as “static systems upgraded annually” and start planning for continuous change.

The recent integration of GPT-5 into Copilot Studio and select Microsoft 365 Copilot environments underscores this shift. Announced as a “meaningful step forward in multimodal reasoning,” GPT-5 enables agents to handle multi-step workflows with fewer errors and greater autonomy. For HR teams, that translates into agents that can schedule interviews, answer benefits questions, and even assist in compliance audits—all without human hand-holding.

But velocity cuts both ways. If models improve faster than training programs can adapt, the risk of shadow AI skyrockets. Employees turn to consumer ChatGPT or standalone agents, bypassing enterprise security. Brown’s advice: provide a sanctioned, secure experience that matches the pace of innovation, or watch shadow IT explode.

The Culture Imperative: More Than a Buzzword

Technology alone won’t deliver benefits if people are fearful or unsupported. Brown’s cultural checklist cuts through the fluff:

• Leadership role modelling. Adoption metrics mirror how visibly senior leaders use and showcase agents.
• Experimentation rituals. Hackweeks, L&D sprints, and AI task forces that normalize trial and error.
• Protected learning time. Without schedule protection, reskilling becomes a checkbox exercise.
• Psychological safety. Mistakes must be treated as learning signals, not career-limiting moves.

“Look at how tools like Microsoft Copilot are being embraced,” Brown says. “Are people experimenting and finding productivity value, or are they threatened and not using it?” When leaders role-model use and actively encourage exploration, adoption metrics rise. When they don’t, the tool gathers dust—no matter how powerful the underlying model.

EY itself has rolled out AI badges, curricula, and enterprise-wide learning pathways. The signal is clear: the organization values AI fluency, and employees who invest time in learning will be recognized. That cultural scaffolding is what turns a potentially disruptive technology into a shared resource.

Governance and Compliance: Non-Negotiables

Agentic AI expands the surface area for privacy breaches, bias, and auditability nightmares. Brown’s team—and countless best-practice sources—recommend an integrated governance stack:

• Data classification + DLP rules that prevent PII from leaking to public LLMs.
• Logging and provenance for every agent decision, so auditors can reconstruct “why.”
• Human-in-the-loop gates for consequential outcomes: hiring, promotion, termination.
• Bias/fairness testing and independent audits for any model used in HR processes.

The regulatory environment is sprinting to catch up. The EU AI Act classifies personnel systems as high-risk, requiring documentation, testing, and human oversight. In the U.S., the EEOC is scrutinizing employment-related AI for civil-rights compliance. For HR leaders, treating compliance as a design constraint—not an afterthought—is the only viable path.

Brown underscores that governance isn’t just about avoiding fines; it’s about preserving trust. Employees who know an algorithm won’t secretly decide their career are more willing to engage with agents.

Skills at the Task Level: HR’s New Lens

Traditional workforce planning looks at job titles. Brown flips that to task-level audits. “Understand which skills AI will impact, which remain uniquely human, and what new roles get created,” he says. That means mapping three buckets:

1. Tasks likely to be automated or accelerated by AI.
2. Uniquely human tasks—empathy, complex judgment, stakeholder relationships.
3. Newly created tasks—prompt engineering, agent orchestration, model-audit roles.

The WEF’s net-positive jobs projection reinforces this approach. Roles won’t vanish en masse; their content will shift. HR must design reskilling and mobility pathways that match the speed of change. Brown’s practical steps:

• Run a 30–60 day task audit focusing on high-frequency, low-risk HR activities.
• Build career maps showing lateral redeployment options before considering layoffs.
• Incentivize reskilling with time budgets and internal credentials tied to performance reviews.
• Launch micro-credentials that signal in-demand skills.

Measuring Success: Brown’s Three Loops

Brown’s investment framework is elegantly simple—three loops that connect pilots to business outcomes:

1. Loop 1 – Do current tasks cheaper, faster, better. Deploy copilots for high-frequency admin work and track time saved, error rate, and employee NPS.
2. Loop 2 – Realize new value. Use agents to unlock new client insights, faster product launches, or broader services. Measure new revenue, customer satisfaction, or speed of decision.
3. Loop 3 – Reinvent the business. When agents are ubiquitous, redesign operating models. This is a strategic bet, not a quick win, and requires separate funding and governance.

“Tie back to why it was implemented,” Brown says. Same output but cheaper? That’s a Loop 1 win. New capabilities—like his third-party risk team using agents for deeper supplier analysis—fall into Loop 2 and 3. Without these metrics, HR teams risk celebrating vanity adoption numbers instead of real business impact.

A Practical Playbook for HR Leaders

Brown’s insights translate into concrete steps that any enterprise can take this quarter:

1. Convene a governance task force (HR, IT, legal, security, employee reps) and publish a charter within 30 days.
2. Complete a rapid task audit and pick 2–3 low-risk pilot use cases in 60 days.
3. Deploy a sanctioned enterprise agent (e.g., Microsoft 365 Copilot with DLP and logging) while opening a sandbox for innovation.
4. Launch a foundational AI badge curriculum for HR and managers; require completion for pilot owners.
5. Instrument KPIs mapped to the three loops and share a monthly dashboard with the executive team.
6. Gate escalation: require human sign-off on any agent output affecting pay, hiring, termination, or legal status; build appeal routes and audit trails.
7. Publish an “AI at work” policy clarifying permitted uses, data handling, and employee rights.

Investment Priorities: Where to Put Your Money

Using the three-loop lens:

• Short bets (Loop 1): Modest investment in copilots for high-frequency admin tasks. Expect swift payback if instrumented correctly.
• Medium bets (Loop 2): Fund cross-functional pilots that create new client value. Plan for longer validation and higher governance costs.
• Large bets (Loop 3): Allocate strategic funds only when agents are widely adopted across comparable firms or when competitive parity is at stake.

Capital commitments signal seriousness but don’t guarantee outcomes. Measure ROI in cost reduction and value delivered—new services, customer satisfaction, decision speed.

Risks and Blind Spots: What Could Go Wrong

Brown and other experts flag recurring pitfalls:

• Executive knowledge gaps caused by procurement delays. Leaders who don’t use the tools can’t lead transformation.
• Treating AI as a static product. This breaks change management. Think continuous improvement, not annual upgrades.
• Shadow AI. If enterprise tools are slow or clunky, employees will find their own—unsecurely.
• Overreliance and deskilling. If managers lean on agents without oversight, institutional judgment erodes. Protect learning objectives that keep human skills sharp.
• Model updates changing behavior overnight. Gate upgrades with staged rollouts and pre-production testing.

A risk matrix for HR:

The Bottom Line

Simon Brown’s message is a clarion call for HR leaders: agentic AI is not a “nice to have.” It is a platform that will evolve in months, not years. The upside—routine cognitive load shifted to agents, people freed for higher-value work—is real. The downside—a workforce that fears technology, broken governance, and runaway shadow IT—is preventable.

For leaders who build the right culture, governance, and learning pathways now, agentic AI will be a tool that lifts people and organizations. For those who don’t, the technology will still arrive—but they’ll be reacting rather than shaping their future.

As Brown puts it: “I’m amazed daily by what I achieve using AI and agents. ChatGPT-5’s recent capabilities are mind-blowing.” The question is whether your organization is ready to harness that amazement—or be steamrolled by it.