The impact of AI on business is accelerating: a radically different future of work and business is being built today. Deloitte’s APEC CEO survey (to be released at the APEC CEO Summit in October) shows that businesses across the region are moving past experimentation and into production. Currently, seven in ten businesses have implemented AI into two or more business functions, and almost a quarter across four or more functions. While today most impact is seen in technology, marketing, sales and service functions, leaders expect AI to make more impact in core business processes and management functions within the next few years¹. 

In this evolving landscape, the questions I hear most from leaders are: How do I prepare our people and processes for AI-driven work? How do I shift from efficiency and cost-saving to drive growth and innovation? The answer is more than technology; it requires collaboration across the entire business.

Enter the agents: Your new digital workforce

Driving this change is the emergence of agentic AI systems, autonomous agents that can reason, remember, learn, plan and act independently across business functions. This is a shift beyond generative AI applications – and one with the potential to build AI-fueled organisations.

Unlike traditional software deployments with rule-based outputs, AI agents allow businesses to scale operations, tackle complex challenges and respond to market shifts in real-time. This offers increased efficiency and flexibility, but also new challenges.

The agentic challenge: Cascading autonomous decisions

Agentic AI creates both opportunity and risk. Organisations must rethink how they manage governance, compliance and collaboration. Truly autonomous systems demand cross-functional orchestration from the outset. It requires deep process understanding, data and engineering skills – and critically, an AI-fluent understanding of the legal, compliance and risk impacts.

Deploying agentic AI requires fundamentally different architecture decisions, and a holistic approach across technology and business. Leaders must define how agents interact with each other and with people, set boundaries for decision making, and ensure fairness and security. 

The challenge extends to what we call “Trustworthy AI”. When agents independently modify business processes, new dimensions of fairness, explainability, accountability, privacy and security become exponentially more complex. For example, an autonomous procurement agent must make cost-effective decisions while operating transparently and without introducing supplier selection bias.

Deloitte’s Path to Scale framework highlights the need for businesses to address interconnected priorities and design for the ethical, operational and strategic demands posed by these new technologies. Increasingly, we at Deloitte support clients to access robust agents and accelerate scalable and trusted deployments to make AI work for them effectively and responsibly.

For forward-thinking businesses, agentic AI is not just about gaining a technical edge; it’s about navigating a foundational shift that will define the future of work, leadership and value creation.

Human-Agent collaboration

For many businesses, the future of work will blend a human and digital workforce. Beyond AI-literacy, employees need new skills for effective human-agent collaboration. And for work design, organisations need to plan for working at two-speeds, human and AI, and increasingly a third dynamic: agentic AI capabilities that are evolving faster than organisations can adapt.

Employees need to monitor autonomous decision-making for bias, understand when agent explanations suffice or when human judgement is needed, and take responsibility for agent actions within their domain. If an AI agent makes biased hiring recommendations or compromises data privacy, how do human collaborators maintain accountability while preserving agent efficiency?

The competitive advantage

Scaling agentic AI isn’t just technology implementation. It’s organizational evolution towards a collaborative human-agent workforce. Today, 44% of businesses across APEC economies assign AI strategy to IT leadership². However, success requires broader leadership and cross-functional coordination. 

Winning with AI means building ‘collaborative intelligence’: the capacity to seamlessly coordinate human expertise with AI capabilities. This calls for strong leadership, broad AI-fluency, and a cross-functional approach across business, technology, talent, legal and risk. 

For leaders building AI-fueled businesses, the answers to trustworthy implementation, value creation and adaptation lie in coordination across business functions.  Collaborative muscle memory is a source of competitive advantage as agentic AI systems become both mainstream and more sophisticated.