Deloitte Survey Shows Most Enterprises Struggle to Scale AI Agents Beyond Pilot Programs

American enterprises face increasing pressure to transition artificial intelligence agents from isolated experimental projects into full-scale production systems, but structural and organizational hurdles continue to slow adoption. According to a recent survey conducted by consulting firm Deloitte and first reported by ZDNET, corporate leaders are struggling to craft integrated, actionable strategies for adopting agentic technology. The research highlights a growing gap between corporate ambition for autonomous software agents and the real-world operational readiness of enterprise infrastructure and personnel.

The Deloitte study, which surveyed 501 senior executive decision-makers managing AI strategies or implementations, reveals that enterprise deployment remains largely fragmented across major business functions. While 42% of surveyed organizations are currently testing small clusters of agents and 43% are expanding single-function deployments, only 15% have successfully deployed fully scaled, orchestrated multi-agent systems across customer support, IT operations, and engineering departments. Despite these early stage rollouts, nearly two-thirds of executives report reevaluating their overarching business models in response to agentic developments, though only 50% state they have a clear image of their future agentic operating framework.

Technical and governance roadblocks represent the primary hurdles preventing companies from moving beyond limited pilot programs. Deloitte found that 72% of respondents point to the lack of a consolidated and accessible data foundation as a primary obstacle to scaling agents. Additionally, 70% of business leaders cited difficulties establishing trust and oversight over autonomous agents, while 67% listed the high financial expense and structural complexity of system integration as critical adoption barriers.

Organizational preparedness remains low across key business capabilities required to support agentic AI environments. Only 36% of executives reported that their company's corporate vision and strategic planning are fully prepared for agentic integration. Furthermore, just 34% of respondents stated their technology infrastructure is ready, followed by 32% for data foundations and 26% for risk management, security, and governance frameworks. The lowest area of preparation lies in workforce readiness, with only 20% of leaders indicating their staff possesses the skills required to work alongside autonomous agents.

While just 16% of executives believe their existing operational workflows are ready for immediate agentic deployment today, respondents project massive process overhauls over the coming decade. Approximately 31% of leaders expect their companies to rebuild business workflows around AI agents by 2028, a figure that surges to 74% by 2030, at which point executives anticipate nearly half of all corporate processes will be redesigned for agentic software. Moreover, 61% of business leaders predict that the majority of these workflows will operate with complete or near-complete autonomy, requiring minimal human intervention. To bridge the present gap, many enterprises are layering agents onto existing legacy software, though Deloitte notes this approach must serve as a temporary mechanism rather than a long-term operational substitute for full process redesign.

Workforce transformation represents another critical area where corporate funding lags behind strategic expectations. Half of the surveyed executives admitted their organizations are failing to allocate sufficient capital toward retraining employees for an agentic workplace. This funding deficit comes as 43% of leaders project substantial workforce disruption resulting from the automation of routine and structured tasks. In response, 71% of surveyed companies report developing basic AI agent literacy programs, while 65% are establishing specific upskilling and reskilling initiatives for positions directly impacted by agentic automation.

Beyond personnel retraining, corporate finance teams face mounting budget demands driven by the operational overhead of agentic software. Executives must navigate rising computational expenses, including token consumption costs and infrastructure investments, alongside the capital required to re-architect business workflows. Deloitte emphasizes that building a functional agentic enterprise requires leaders to look beyond technical upgrades and implement broader organizational changes, such as recalibrating corporate metrics around agent usage patterns like token execution, reallocating human capital toward non-deterministic work, and establishing new collaborative operational models between human staff and autonomous software.