The ROI Honeymoon Is Over: AI Must Now Prove Its Worth

Over the past two years, companies have dived headfirst into artificial intelligence, spurred by the hype and potential of generative AI (GenAI). However, the honeymoon phase is quickly fading. Experimentation is no longer sufficient, and businesses must now demonstrate concrete returns on their AI investments. With over 70% of AI projects stuck in pilot phases and failing to advance, companies are feeling the pressure to turn trials into tangible outcomes within the next 12 months.

This shift reflects AI’s growing maturity. Early AI adopters could afford to test the waters, but the current climate demands that AI initiatives prove their worth through increased productivity or revenue growth. However, the challenges are significant. As AI transitions from a buzzword to a business imperative, I’ve spoken with many organizations (and their consultants) that are beginning to realize the operational and cultural shifts needed to scale AI successfully.

We Are Unprepared for What’s Coming

GenAI is set to disrupt a wide array of cognitive, nonroutine tasks in fields ranging from finance and health care to education and law. According to a recent Brookings Institution report, more than 30% of workers could see over 50% of their job tasks impacted by AI. The report emphasizes the urgent need for new policies, mental models, and worker empowerment mechanisms to mitigate risks and capitalize on opportunities.

Despite the high stakes, Brookings argues, we are unprepared for this transformation. Policymakers and industries have been slow to address potential disruptions to the workforce. The current lack of urgency is concerning, especially as businesses rush to implement AI without considering its long-term impact on workers. We lack policies that address how AI will affect worker protections, labor markets, and overall economic inequalities.

The Scale of the Challenge: Cognitive Disruption

Unlike past waves of automation, GenAI threatens higher-paid, cognitively complex roles. Sectors such as health care and education—fields traditionally seen as safe from automation—are now at risk. Brookings research shows that over 50% of tasks in these fields could be disrupted by AI, impacting millions of workers. Jobs that rely on problem-solving, decision-making, and creativity are now exposed to significant AI-driven automation.

This shift brings several unknowns:

1. The extent to which AI will replace or augment human labor.
2. Whether AI will boost productivity in cognitive roles or merely reduce human involvement.
3. The potential exacerbation of income inequality.
4. The ability of the current workforce to adapt quickly enough to avoid widespread displacement.

Regulatory Gaps and a Lack of Worker Empowerment

The Brookings report highlights a critical gap: the lack of worker empowerment in AI design and implementation. While companies are racing to adopt AI, few are involving their employees in these decisions. This is a missed opportunity, as workers could provide valuable insights into how AI could complement their roles.

Additionally, there are few regulatory frameworks to guide companies on ethical and responsible AI implementation. Without clear guidelines, the risk of AI exacerbating inequality and deepening the divide between high- and low-skilled workers grows. Brookings stresses the need for policies that ensure the equitable distribution of AI benefits and worker protections.

What We Don’t Know: The Big Questions

There are several unknowns regarding AI’s impact on the workforce, including:

1. How quickly AI will advance from augmenting human labor to replacing it.
2. Which industries and workers will benefit or suffer the most.
3. How these changes will affect income inequality.
4. Whether AI can boost productivity without displacing jobs.
5. What strategies are necessary to protect workers and ensure the equitable distribution of AI benefits.

The lack of clarity on these questions complicates efforts to prepare for AI’s impact. While AI has the potential to boost productivity, it also risks devaluing jobs, reducing wages, and increasing worker precarity, particularly in cognitively intensive fields.

A Few Are Starting to Understand the Bigger Picture

While many companies remain in the experimental phase of AI, a small percentage are beginning to realize that experimentation is only part of the equation. According to the Upwork Research Institute, 27% of companies, known as Work Innovators, are thriving by integrating flexible work, distributed teams, and advanced technology into a unified system. These companies prioritize worker engagement, clear communication, and seamless integration of technology into their operations.

The key takeaway from Upwork’s research is that innovation doesn’t require massive spending or risky bets. The companies that succeed are those that thoughtfully integrate new technologies with a focus on employee empowerment and engagement, navigating the AI transition with the agility needed to thrive in an uncertain world.

Learning and Development: A Transformative Opportunity

One of the most promising areas for AI integration is learning and development (L&D). Chief learning officers (CLOs) are beginning to embrace AI as a tool for innovation, particularly in upskilling employees and enhancing personalized learning experiences. For example, Udemy’s CLO leveraged GenAI to streamline assessment creation, enabling managers to develop practical skills through real-time simulations and feedback.

This application of AI goes beyond basic task automation. It allows for the kind of personalized, adaptive learning that traditional methods could never achieve. In a culture that promotes growth and experimentation, AI can be a powerful tool for human transformation. The potential to tailor learning experiences to individual needs and preferences represents a breakthrough in how organizations approach employee development.

In the next year, we will likely see a transformation in how organizations approach L&D. As AI continues to mature, forward-thinking CLOs will lead the charge in using AI to create more dynamic, personalized learning environments that empower employees. These early adopters will demonstrate the true value of AI by showing that not only can it improve efficiency but also foster human growth and development.

Urgency Needed in Policy and Business Practices

The Brookings report calls for urgent action from both policymakers and business leaders. We need to develop frameworks that address not only AI’s technical capabilities but also its broader implications for the workforce. Doing so will include:

1. Creating ethical guidelines for AI implementation.
2. Involving workers in AI decision-making.
3. Ensuring that AI-driven productivity gains are shared across the workforce.

Moreover, we need to shift our mental models about work and AI. Instead of seeing AI purely as a cost-cutting tool, we should focus on how it can empower workers and enhance their capabilities. This means not only using AI to automate tasks but rethinking how AI can augment human roles, leading to more meaningful work and higher productivity.

Conclusion: The Stakes Are High

The challenges posed by GenAI are significant, and we are not adequately prepared to address them. Without urgent action, AI’s potential to disrupt middle- and high-income jobs could lead to increased inequality and job insecurity for millions of workers. Policymakers and business leaders must act now to ensure that AI’s benefits are shared equitably and that workers are empowered to shape AI’s role in the future of work.

Expanding worker engagement, developing robust policies, and reframing how we think about AI’s impact on jobs are critical steps in preparing for the changes ahead. Only by addressing these challenges head-on can we ensure that GenAI enhances, rather than undermines, the future of work.

The pressure to prove AI’s return on investment is mounting, but we are largely not yet ready for the organizational and workforce changes that must accompany AI’s implementation. While companies such as W.W. Grainger and Palantir are focusing on getting AI projects to deliver results within a year, the reality is that many organizations are struggling to move beyond experimentation. The critical missing link is the recognition that AI’s success hinges not just on the technology itself, but also on how well it integrates with the human side of the business.

The honeymoon phase of AI is over. Now it’s time for companies to prove that AI can deliver real value—not just in terms of productivity and revenue, but in the way it transforms how we work, learn, and grow as individuals and as organizations.