There is no shortage of published thought leadership on the potential of artificial intelligence. But what often gets overlooked is the gap between theoretical potential and the reality of deploying it successfully inside a global enterprise. Gallup’s 2026 State of the Workplace report reveals a stark disconnect: while 65% of employees say AI has boosted their individual productivity, only 12% strongly agree the technology has actually transformed how work gets done in their organization.
Having led tech transformations from both sides — as an executive driving innovation at Canadian National Rail and as a partner accelerating AI outcomes for customers with Tenarai — I’ve learned that successful digital transformation thrives on people. In the agentic era, the unlock for enterprise AI outcomes is empowering employees to unlearn the tools and habits they have relied on for years, helping them embrace new capabilities with confidence.
As the industry shifts from passive AI assistants to autonomous AI agents, traditional software delivery models fall short because they overlook the human-AI connection. Bridging this gap requires elevating change management into a foundational pillar of enterprise AI transformation, aligning tech with the context and culture of the enterprise.
From assistants to active agents
To understand why legacy engineering structures fall short on scaling AI outcomes, let’s look at how technology changes human workflows. Generative AI serves primarily as a task-based drafting tool, keeping humans comfortably in the driver’s seat to review and approve.
Agentic AI, in contrast, is goal-driven, context-aware, and collaborative; acting more like an autonomous digital co-worker. This shift is already reshaping complex enterprise operations, but its success hinges on people’s ability to adapt.
A recent data migration we led for a global logistics leader illustrates this challenge. While moving its data pipeline from a legacy SAS architecture to the cloud presented significant engineering complexity, the greater challenge was actually guiding its long-tenured business analysts in unlearning many years of established workflows.
This tension reveals the fundamental flaw in traditional AI adoption. When employees transition from controlling task-based tools to supervising autonomous agents, a vulnerability gap is exposed. Legacy IT structures, often operating in silos, are inherently built to fail in this environment, leaving frontline staff stranded between old habits and new autonomous capabilities. Bridging this divide requires installing a modern delivery model engineered around human-centered collaboration.
Architects of trust: FDEs and agentic design leads
To close this adoption gap and scale agentic AI from proof-of-concept experiments to real business outcomes, organizations must deploy a dual-role framework. This model embeds change management, domain expertise, and operational empathy directly into the technical pipeline through two key roles: the Forward Deployed Engineer and the Agentic Design Lead.
Traditional software teams frequently write code in silos based on slow-moving specifications. The Forward Deployed Engineer (FDE) model eliminates this bottleneck. FDEs operate as adaptable technical architects embedded directly at the intersection of business processes, corporate context, and live production code.
Working within cross-functional units, FDEs collaborate closely with specialized AI agents to compress build cycles from months into weeks. By working alongside business users, FDEs rapidly translate operational goals into scalable, production-ready systems while resolving friction in real time.
While the FDE drives technical execution, the Agentic Design Lead focuses entirely on the human-AI connection and organizational alignment. This specialized role reshapes business workflows around agentic capabilities before development begins, securing active adoption at every stage. By analyzing user behavior, the Agentic Design Lead remodels the design process with the actual users in mind, ensuring a seamless human-AI connection is pulled all the way through the final solution.
Managing the risk of unlearning
Guardrails and governance are essential because unlearning decades of defined software habits can be highly uncomfortable for an organization. To scale agentic workflows safely, leaders must establish a clear map of where unlearning is encouraged and where it is not, dividing enterprise operations into two distinct risk categories: deterministic and probabilistic.
Deterministic controls govern mission-critical environments where errors risk physical operations or human safety; such as warehouse operations, heavy rail dispatching, air traffic management, or industrial manufacturing. These high-consequence domains demand rigid, non-autonomous safeguards.
Conversely, probabilistic frameworks support domains where autonomous agents excel, including demand forecasting, targeted marketing, and automated document analysis. For these adaptive environments, leaders should establish secure sandboxes where cross-functional teams can safely experiment, evaluate performance logs, and scale proven use cases incrementally.
Change management as an internal marketing campaign
Even the most advanced AI architecture fails if the workforce resists adoption. Enterprise transformation cannot succeed through executive mandates or static dashboards. Leaders must execute it like an internal marketing campaign, earning employee buy-in across the organization and purposely guiding teams through grassroots engagement.
Organizations can build momentum through interactive workshops, innovation days, tailored training, and immersive boot camps that demystify new tools through hands-on exposure. Meanwhile, low-pressure forums such as team connection meetings or informal brown-bag sessions provide safe spaces to address concerns, establish trust, and cultivate creative confidence across teams. We’ve done all of these activities and though some may seem simple, they work.
Winning in the agentic era isn’t about outspending competitors on models, it’s about understanding people and navigating them through change. Operational success belongs to the organizations that view change not as a technical hurdle, but as their most important product. By guiding their employees in the art of unlearning, leaders transform workforce friction into adoption, ensuring that as agents take on the heavy lifting, human potential remains the true engine of innovation.
We’ve featured the best AI website builder.
This article was produced as part ofTechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here:https://www.techradar.com/pro/perspectives-how-to-submit
