It has been nearly three years since the first release of OpenAI’s ChatGPT, and the prevailing narrative is that professionals not yet integrating AI into their work risk being left behind. Executives and thought leaders often frame AI as an inevitability – an unstoppable wave of transformation sweeping through every industry.
On the ground, however, the picture is more complicated. The employees tasked with actually using AI tools are often the most skeptical of their usefulness and effectiveness. Their hesitation doesn’t stem from ignorance or stubbornness, but from an intimate understanding of the messy, high-stakes realities that AI is meant to reshape if it lives up to the hype and investments made.
When I transitioned from the technology sector into the environmental space, I expected to find advanced digital systems guiding compliance, sustainability reporting, and facility operations. Instead, however, I was struck by the sheer amount of manual and paper-based work that still defined the day-to-day. Even more surprising was the reluctance of frontline staff to adopt more technology-forward processes, despite clear efficiency gains.
This tension became especially visible during customer onboarding and training. At my workplace, the environmental compliance software company where I work, we often encounter resistance not only to digital record-keeping but also to the experimental AI features I am implementing with industrial companies across North America. The pattern has become clear: the primary challenge is not the technical leap itself, but the cultural and structural transformation required inside organizations. Getting people to adapt to new, AI-driven ways of working is less about algorithms and more about shifting workflows, incentives, and mindsets. In many ways it is the similar roadblocks faced in any digital transformation, but AI is moving at a far faster pace and with potentially bigger economic impacts than other digital transformations. The speed at which we are asking workers, leaders, and organizations to adapt is both amazing and troubling for many.
The Shifting Nature of Compliance Roles in the AI Era
AI is changing not just what work gets done, but who does it, how it’s done, and where effort is focused. According to a recent McKinsey report, nearly all companies are investing in AI, but fewer than 2% believe they have reached full maturity in integrating AI into their workflows. Four major shifts are becoming visible:
1. From Repetition to Oversight
Traditional roles in environmental compliance weighted toward manual data entry, filing, auditing, checking compliance reports – these are increasingly being automated. The human value moves to oversight, exception handling, verification, and interpretation. In environmental compliance, for example, AI tools can flag anomalies in waste classification or predict where facilities may fall short of regulatory requirements. Humans still make the calls, especially where nuance and risk are high.
2. Skillset Elevation and Hybrid Roles
As automation takes over low-complexity work, the demand for hybrid skills rises: domain expertise + digital literacy + data interpretation. Workers need to understand both their environmental or regulatory domain and how AI systems produce outputs. They must also judge when AI outputs are trustworthy and when human judgment should override them.
3. Cultural and Organizational Change
Adopting AI is rarely about plugging in a tool, it’s about changing how people think and how teams cooperate. Organizations that expect AI alone to deliver transformation often underinvest in the human side of change, neglecting the need for training and redefinition of roles. Frequently, there is a lot of resistance from the teams, not because the tool fails technically, but because people feel unsure, unprepared, or threatened.
4. Leadership’s Expanded Role
Finally, for executives, the question is no longer just “What AI should we adopt?” but “How will we lead people through this change?” Leaders must anticipate bottlenecks in adoption, prioritize explainability, and build feedback loops that surface real user concerns early. They also must think of AI deployment as a people strategy as much as a technology strategy.
Case Study: Waste Tracking and the Shift from Data Entry to Data Review
To ground these ideas, here’s how workforce transformation has played out in practice through one AI project I worked on in environmental compliance.
A manufacturing customer of ours generates hazardous waste that must be shipped off-site for proper handling. This requires a regulatory document known as the hazardous waste manifest. When waste leaves the facility, a paper manifest is produced containing key information about the waste type, quantity, and origin. One copy stays with the generator, and another travels with the waste until disposal.
This single-copy, paper-based system creates a host of problems:
- Manifests can be lost or misinterpreted.
- Data is often miscalculated or misreported.
- At year’s end, staff must manually re-enter manifest data into spreadsheets to meet reporting requirements – a process that can take several full days of work for one facility.
Faced with this, our question was simple: could AI handle the paperwork?
The Proposed Solution
The idea was straightforward: allow users to upload a photo of a manifest, apply AI models to extract the key information, and then feed that data into dashboards that could automate reporting in the required regulatory format. In theory, this would reduce errors, free up staff time, and provide real-time insights into waste streams.
The Reality of Implementation
In practice, things were far less seamless. The structure of waste manifests is highly specific, and training a model that could reliably interpret them took time. Early results were uneven, leading to friction with customers and, more importantly, distrust. Many of these facility teams had little exposure to AI in their daily work and were reluctant to rely on outputs they couldn’t verify. When the model produced errors, confidence eroded quickly.
As performance improved, we added safeguards such as validation limits on reported values. Over time, customers began to see the value: fewer hours lost to manual entry and clearer visibility into waste streams. But adoption remained cautious, underscoring how cultural trust is as important as technical accuracy.
Building Trust Through Redundancy
One breakthrough was connecting our system to the EPA’s e-Manifest API, which provides a centralized digital record of manifests. This aligns with recent EPA Final (Third) Rule updates, published in July 2024, which extend e-Manifest requirements to include export manifests and enhanced manifest-error correction procedures. This gave facilities a safety net: if our AI missed values or introduced discrepancies, the EPA data served as a secondary validation layer. In practice, this meant customers were alerted if the AI-extracted data did not align with federal records – either signaling an error in our model’s interpretation or highlighting that incorrect information had been reported upstream to the EPA.
This additional layer of transparency helped overcome skepticism. Workers were no longer forced to blindly trust the AI; instead, they became supervisors of its performance, empowered to reconcile mismatches and escalate issues.
The Workforce Shift
The most significant transformation wasn’t technical but human. Previously, compliance staff spent days manually transcribing numbers from paper forms into spreadsheets. With AI and EPA integrations, their role shifted from data entry to data review and exception handling. Instead of being clerks retyping information, they became analysts ensuring accuracy, consistency, and regulatory compliance.
This case highlights the broader workforce trend: AI does not eliminate the work – it changes its nature. For environmental compliance officers, the shift is from repetitive manual processes toward higher-value oversight and decision-making.
Beyond Environmental Compliance: A Universal Shift
Similar transformations are visible across other industries. In finance, AI systems flag suspicious transactions, shifting analysts’ focus to deeper investigations. In healthcare, AI supports radiologists by surfacing potential anomalies, while doctors concentrate on treatment planning and patient care. In manufacturing, predictive maintenance tools reduce repetitive inspections, freeing technicians to handle more complex repairs.
The pattern is consistent: automation takes over repetitive tasks, while people move toward higher-value oversight, judgment, and strategy.
Lessons Learned: What I’d Keep in Mind Next Time
Looking back on this project, if I were guiding a team through AI adoption again, here are a few things I’d keep in mind:
- Start small. Begin with a limited workflow or dataset to prove value before scaling.
- Build trust early. Share error rates, show how outputs are validated, and be transparent about limitations.
- Keep humans in the loop. Make it clear that AI is a co-pilot, not a replacement.
- Invest in training. Help people learn not only how to use the tool but how to interpret and question its outputs.
- Celebrate small wins. Even modest time savings or accuracy improvements build momentum for adoption.
AI is not simply a technology issue, it’s a people issue. Across industries, automation and AI tools are shifting roles rather than eliminating them outright. What defines successful organizations will be their ability to manage the human side of transformation: changing workflows, redefining roles, building trust, and investing in skills.
For operational and compliance leaders in manufacturing companies – and beyond – the question is not whether to adopt AI, but how to do so in a way that empowers the workforce rather than disenfranchises it. The future of AI is not cold algorithms replacing human effort, it is humans and AI working in concert, each doing what they do best.
