Most enterprises say they want to deploy agentic AI within three years, but the vast majority lack the infrastructure to make it work. 85% of organizations plan to adopt AI agents by 2029, but 76% report their current operations, people, and workflows can't support that change, according to research published by MIT Technology Review and enterprise AI platform Ema. The disconnect stems from a structural problem: companies are layering AI agents onto legacy operations instead of redesigning the organization around them.
AI agents deployed at scale can accelerate business processes by 30% to 50% and reduce low-value work time by 25% to 40%, early proving grounds in customer service, HR, and sales show. But capturing that value requires rethinking three core pillars: the technology stack, workforce structure, and success metrics. Ema and HFS Research coined the term agentic business transformation (ABT) to describe this shift, arguing that existing vocabulary around digital transformation, AI transformation, and copilots fails to capture the systems-level change required.
“They're embedding AI employees into what is a human operating model”— Prasun Shah, global CTO for workforce consulting at PwC UK Consulting
The technology problem is architectural. Legacy tech stacks were designed for human-operated, application-centric workflows with linear processes and discrete steps. AI agents function as connective tissue, moving across layers to coordinate high-level tasks and retrieve data from multiple systems simultaneously at machine speed. That capability creates competitive differentiation, but only when enterprises surface access to multiple datasets and applications so agents can develop tacit knowledge and contextualize decisions.
Key facts
- 0185% of organizations say they want to deploy agentic AI within the next three years, but 76% report their current operations and infrastructure can't support that change.
- 02Early deployments show AI agents can accelerate business processes by 30% to 50% and reduce low-value work time by 25% to 40% when deployed at scale.
- 03McKinsey predicts three-quarters of current jobs will require redesign, upskilling, or redeployment by 2030 as agentic AI reshapes workforce structures.
- 04One enterprise customer tripled its measured ROI from agentic AI within two quarters after overhauling metrics from tool-level outputs to business outcomes.
- 05An AI agent can handle 1,000 customer interactions in the time it takes a human to handle 10, rendering traditional activity metrics obsolete.
Workforce structures will fracture under the same pressure. The hierarchical model inherited from early industrialization assumes managers coordinate execution, employees progress by optimizing output from teams below them, and tasks are delineated by strategic business units. AI agents that execute, coordinate, and optimize tasks without managerial oversight blur those lines. Managers freed from execution-based tasks will instead manage trust, explainability, psychological safety, and status dynamics in hybrid teams. McKinsey predicts three-quarters of current jobs will require redesign, upskilling, or redeployment by 2030.
Traditional workforce metrics collapse when AI agents assume ownership of core processes. Activity metrics like calls handled or reports filed become meaningless or actively misleading when an AI agent can process volume at machine speed. One of Ema's large enterprise customers tripled its measured ROI from agentic AI within two quarters after switching from tool metrics like cost per query and AI accuracy to outcome metrics like the percentage of contracts reviewed without human escalation.
The metric overhaul forced the company to stop building point solutions in high-volume, low-complexity workflows and start deploying AI employees where outcome value was highest. Integrating outcome-based metrics requires reconfiguring reward structures, talent management processes, and accountability frameworks. Ethical and fiduciary responsibilities will remain with human employees, but operational accountability will diffuse across human-AI teams.
“When you add AI employees into the workforce, activity metrics become meaningless or actively misleading”— Surojit Chatterjee, CEO of Ema
The accountability question raises unresolved lines of inquiry for senior leadership. Who is accountable when an AI employee makes a mistake? What happens when AI and humans disagree? What guardrails protect customers? These questions don't have settled answers yet, but enterprises that start addressing them now will close the gap between their stated ambition and their actual execution capability.
“An AI employee can handle a thousand customer interactions in the time it takes a human to handle ten. If you measure success by interactions handled, you'll conclude the AI is working brilliantly while missing whether any of those interactions actually drove customer satisfaction, retention, or revenue.”— Surojit Chatterjee, CEO of Ema
The gap is wide. 76% of organizations report they're not ready for agentic AI across people, processes, and workflows, even as 85% say they plan to deploy it within three years. The mismatch suggests most enterprises are treating AI agents as productivity aids or point tools rather than as participants in value creation that require systems-level change. Early movers that redesign operations around agents are seeing material gains; laggards layering agents onto legacy operations are seeing marginal improvements at best.
Systems-level change is gradual, but the dialogue has to start now. Enterprises that begin internal conversations around workforce redesign, technology stack adaptation, and outcome-based metrics will be positioned to extract real value from agentic AI. Those that don't will find themselves in the 76% — ambitious but unprepared, deploying agents into structures that can't support them and wondering why the ROI never materializes.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
