How to ensure AI agents become the strategic partners in your business to unlock AI’s full value

The next frontier of business isn’t just about faster technology. We’re witnessing a fundamental shift in how value is created through artificial intelligence (AI) agents that operate as strategic partners in decision-making.

The most advanced act as orchestrators – coordinating intelligence across systems and paving the way for persona-centric super agents. Together, they unlock new business models, enable cross-industry ecosystems and drive large-scale transformation.

CEOs are taking notice. According to our latest Global CEO Outlook, 57% expect AI agents to significantly impact their organizations, alongside generative AI. Across our own business and with our clients, we’ve deployed agents in areas ranging from client delivery to internal operations.

Through these experiences, we’ve learned what accelerates adoption, what slows it and what it takes for people and agents to collaborate responsibly and effectively.

As agents increasingly become teammates or collaborators, the lessons from the last three years can guide us into the future, especially as organizations reshape what work looks like, rethink the services and products they deliver and navigate new questions of trust and responsibility.

In 2024, KPMG launched a study to quantify the potential impact of AI on labour productivity for enterprises.

Leveraging our patent-pending AI Value Assessment, the detailed analysis of more than 17 million companies in the report shows that fully embracing agentic AI could unlock approximately $3 trillion in global productivity gains equivalent to a 5% improvement in profitability (earnings before interest, taxes, depreciation and amortization – EBITDA) for the average Fortune 1000 company.

On a macroeconomic scale, widespread use of AI agents could become one of the strongest drivers of new streams of value in the next three to five years. AI agents are joining workforces across industries but long-term value emerges only when people use them as responsible partners.

The conversation is shifting from what agents can do to how they collaborate by reimagining roles, capabilities and cultures.

However, driving enterprise-wide adoption and capturing its full economic potential is not as simple as giving people access to AI tools. Nor is it about automating entire roles.

The real value emerges when organizations evolve from using agents to augment discrete tasks toward re-engineering end-to-end processes. It’s a shift that echoes the 1990s re-engineering revolution.

This means rethinking the entire back office and challenging the notion of functional silos. For early movers, this also represents an opportunity to unlock new levels of creativity and to reshape business models in ways that laggards may struggle to match.

AI only scales safely and sustainably when trust is built in from day one. Trusted organizations attract customers, collaborators and talent while enabling resilience. The question is no longer whether AI is being used but how and whether a given agent can be trusted.

The first wave of agentic AI focused on pilots and prototypes. The next must address deeper system challenges, such as agent sprawl, misalignment and emerging security vulnerabilities and data-protection risks as agents move across internal and external data flows.

Scaling responsibly requires an “agent control system”: a governance layer that oversees how agents are deployed, monitored and evolved.

Building trust at scale spans three dimensions:

Global standards such as the European Union’s AI Act and ISO 42001 are rapidly becoming practical frameworks for embedding trust via assurance into AI agent development and deployment.

Early adopters not only reduce risk but also position themselves to capture value sooner and scale with greater confidence.

AI can generate insights at extraordinary speed but judgment and accountability rest firmly with those who use and oversee it. The workforce’s ability to understand, question and collaborate responsibly with AI will determine whether AI amplifies value or erodes it.

New roles are emerging from “orchestration engineers” who shape how agents think and execute, to “responsible AI/trust engineers” who build the guardrails. Success requires business leaders to clearly define these evolving roles, communicate transparently about the organization's AI journey and invest deliberately in preparing their workforce.

It requires sustained training to build confidence and competence, clear top-down expectations that agents will be used in daily work and deliberate time for employees to explore, experiment and learn.

In the game of value creation, the advantage will belong to those who embrace a human-plus-agent model that is ethical, responsible and bold – and who equip their people with the environment, skills and leadership support needed to realize its full potential. For leaders, there is no sideline – they must personally engage with AI to lead with authenticity.

Employees must also have skin in the game. The incentive for the workforce to embrace AI isn't just about future-proofing their skills. At KPMG, employees who are avid users of AI have been shown to materially outperform their peers. When employees see a direct, measurable link between AI use and personal success, adoption can become a pathway to excellence.

We're helping our clients put this into practice. One enterprise recently partnered with KPMG to create an AI-powered "Career Companion" for 15,000 employees, providing personalized career paths and skill-development plans aligned with company goals.

The result: more than 650,000 skills proactively built and a 99.75% reduction in time required to generate skills and job architectures, demonstrating how AI-enabled workforce transformation accelerates both individual growth and enterprise agility.

The next decade will be defined by the people empowered to collaborate, innovate and solve problems through agents. Early evidence points clearly in this direction.

KPMG is working with academic researchers to generate real-time insights into how early-career professionals engage with agents and which skills matter most. The early findings are clear: digital fluency is no longer enough. Ethical judgment and the ability to collaborate effectively with AI systems will be the true differentiators for next-generation talent.

The goal is not simply to build a more technically skilled workforce. It’s to develop one that is thoughtful, adaptable and grounded in responsible decision-making.

To unlock this future, organizations must design for trust, build for people and invest with courage. In the game of value creation, the advantage will belong to those who embrace an ethical, responsible and bold human-plus-agent model.