The agentic reality check: preparing for a silicon-based workforce

The agentic reality gap: bridging promise and practice 

Enterprises are rapidly exploring agentic AI, with Gartner predicting that 15% of daily work decisions will be made autonomously by agentic AI by 2028, a significant increase from virtually none in 2024. Despite this momentum, many organizations face substantial hurdles in translating agentic pilots into production-ready solutions. Deloitte’s 2025 Emerging Technology Trends study indicates that only 14% of organizations have deployable solutions, and just 11% are actively using these systems in production. 

Several fundamental infrastructure challenges contribute to this gap. Traditional enterprise systems - often not designed for real-time, autonomous agent interactions - create significant bottlenecks, with Gartner forecasting that more than 40% of agentic AI projects will fail by 2027 due to legacy system incompatibility. In addition, data architectures built around batch-based Extract, Transform, Load (ETL) processes introduce friction for agent deployment. Many organizations also attempt to automate existing, human-centric processes instead of reimagining workflows for an agent-native environment, resulting in inefficient implementations and, in some cases, “workslop”- where poorly designed agentic applications actually increase operational burden. 

Redesigning operations for a human - digital future 

Leading organizations are addressing these challenges by adopting a systematic approach to agentic transformation, moving beyond simply layering agents onto legacy workflows. This requires stepping back to examine end-to-end processes and leveraging agents’ distinctive strengths, such as continuous, high-volume task execution without human constraints like breaks or working hours. 

This strategic redesign is already taking shape across industries. At HPE, an enterprise software and services company, an AI agent called Alfred supports internal operational performance reviews. Toyota uses agentic tools to gain real-time visibility into vehicle arrivals at dealerships and is planning to empower agents to resolve supply chain issues - bypassing the need for human interaction with complex mainframe systems. These examples signal a shift away from traditional application modernization toward enabling agents to bridge legacy system gaps directly. 

Perhaps the most profound change is the recognition of agents as a new category of labor - a “silicon-based workforce.” Organizations are increasingly integrating these digital workers with their human, or carbon-based, workforce, enabling people to focus on higher-value activities such as governance, compliance, and growth strategy. Mapfre, an insurance company, for example, deploys AI agents to handle routine administrative tasks in claims management, while keeping humans in the loop for sensitive customer interactions. Biotech company Moderna has gone a step further by merging its technology and HR functions under a Chief People and Digital Technology Officer to better integrate talent and technology. 

Effective implementations also rely on specialized agents orchestrated at scale, automating entire workflows through emerging standards such as the Model Context Protocol (MCP), Agent-to-Agent Protocol (A2A), and Agent Communication Protocol (ACP). Together, these approaches enable a “microservices model for AI,” reducing complexity while supporting scalable orchestration and platform flexibility. At the same time, organizations are introducing FinOps frameworks for AI, essential for monitoring and controlling agent-driven costs, particularly in token-based pricing environments. 

The shift toward agentic AI represents more than a technological evolution; it is a fundamental organizational transformation that will reshape how enterprises operate, compete, and create value. Organizations that master agent-native process design, multi-agent orchestration, and silicon workforce management will be best positioned to thrive in an increasingly automated economy. Ultimately, success will depend on creating new forms of human - AI collaboration that capitalize on the complementary strengths of both human and silicon-based workers. 

To explore how organizations can strategically embrace agentic AI and prepare for the future of work, read the full Deloitte Tech Trends report.