It’s only natural for passions to cool when it comes to any advanced technology. And as we wrap up a full year of the State of Generative AI in the Enterprise survey, we can see that this lifecycle is clearly holding true for Generative AI (GenAI).

Despite the zeitgeist surrounding GenAI—or perhaps because of it—businesses are now entering a more pragmatic stage in which both the benefits of GenAI and the challenges with scaling are becoming clearer. In a recent Deloitte TECHTalks podcast episode, we delved into both sides of the equation—and what they mean to organizations as they work to get the most out of GenAI. Some of our thoughts are shared below.

From speed to scaling

The latest survey suggests that while GenAI is moving at lightning speed, organizations are still moving at the speed of, well, organizations. For most organizations, scaling their GenAI proofs of concept is still mostly a work in progress. Right now, over two-thirds of survey respondents say that 30% or fewer of their current experiments will be fully scaled in the next three to six months.

It’s true that we are seeing an increase in technical preparedness of organizations and that budgets are being dedicated to building the technology infrastructure critical to implementing GenAI. Over the past year, respondents believe their organizations have most improved their GenAI preparedness in the critical areas of technology infrastructure. So, what is slowing down the ability to scale?

Part of it could be a lack of GenAI fluency. C-Suites and boards—generally not as steeped in the technical aspects of GenAI—should be consistently engaged in GenAI adoption, especially when you consider that it’s a multidisciplinary effort involving practically every executive function in an organization. Implementing an ambitious GenAI strategy will most likely require involvement from multiple executives and leaders. But despite a real desire for fluency among executives, it doesn’t appear to be at a level yet where they can actually use their own transformational levers to drive adoption of these technologies throughout the organization.

Another factor is that AI technology is often moving faster than large businesses are changing. The models are progressing even more quickly than expected, exceeding performance thresholds. As such, large, complex companies may struggle to adapt to change. Some executives are expressing energy and optimism around the power of the technology to drive business value, yet some may have frustration over the multitude of experiments and proofs of concept that have not yet scaled.

To speed the move towards scaling, executives should understand and be clear on their GenAI strategy and goals. Is it about productivity, cost effectiveness, growth, innovation, speed? How can you measure value against those objectives? Many business leaders are highly focused on creating new markets, richer ways of delivering value to customers and on the need to leverage tech-driven innovation beyond efficiency plays—extending into use cases that drive top-line growth. Focusing on strategic value from the top down can help clarify next steps and determine which use cases are most important to scale.

The human element

GenAI is talked about everywhere these days in both the business world and society at large—and that awareness can cause “fearers” and “reverers” to emerge within an organization. Fearers are employees who often don’t understand the technology and have heard or read about GenAI replacing some percentage of the workforce (or even replacing humanity itself). And then you have employees that revere GenAI and are eager to use the latest tools—but perhaps don’t use them in a way that's optimal, or perhaps even safe for a business, or doesn’t serve a business need.

Part of addressing this challenge is promoting familiarity with GenAI tools among workers. That means safely giving the technology to as many people in the organization so that they not only understand how to use it but also recognize its limitations and the risks. But despite all the hype, survey results show that access to GenAI tools is still largely limited to less than 40% of the workforce—and for most organizations, fewer than 60% of workers who do have access, use it on a daily basis.

Using these tools can do a lot more than combat fearfulness and misuse of GenAI—and a lack of access could mean organizations are missing out on an added bonus: innovation. Innovation in this space doesn’t necessarily come from the boardroom or from management alone. It can also come from workers experimenting and exploring. Often, this is how organizations generate the best ideas and many companies are finding that this is a great way to drive innovation.

Business leaders are also starting to realize that workers will require significant upskilling and reskilling in order to take full advantage of the GenAI opportunity. Upskilling in core technology will be viewed as table stakes and not sufficient in isolation. Increasingly, talent will most likely need the ability to combine deep technology skills with broader perspectives, ensuring that innovative technologies are deployed in ways that improve quality of life.

The role of risk

In the survey, respondents listed regulation, managing risks, and compliance as some of the top barriers holding them back in developing and deploying GenAI tools and applications at scale. It’s understandable given that the adoption of GenAI can encompass risks that are challenging to identify and quantify given how relatively new it is.

Yet, for an organization seeking to stay competitive, eschewing GenAI is not an option. The question, therefore, should not be between “yes” or “no”—but rather “how” to deploy GenAI in a way that enables the organization to manage risk within their risk appetite. Chief risk officers, compliance officers, or even HR leaders should be supported in these efforts and encouraged to focus on the positive - what they can manage when it comes to GenAI—namely, organizational readiness, especially in areas such as data privacy and confidentiality, risk management, governance, regulatory compliance, and workforce.

Executives and board members assessing risk should dig deeper with GenAI given its startling capabilities and the speed at which it is developing. To have an informed discussion about risks—such as how information is stored or used to train large language models—a more sophisticated level of understanding will likely be required than with previous technological waves.

Getting to the real ROI

While a slower path to scaling is what we’re now seeing—64% of organizations in the survey have fewer than 20 GenAI experiments—the actions are more targeted. That suggests companies are taking the time to test out the capabilities of GenAI, what it can do, how it can benefit the organization, and now explore new developments such as “agentic AI” (52% of survey respondents express interest in these systems that can complete complex tasks with little, or no human intervention).

Those experiments are also focusing more and more on core business areas—moving beyond the perception that GenAI is for personal productivity alone. In fact, three-quarters of the survey’s respondents agree that value is being achieved with the most advanced initiatives. So, a more measured GenAI journey to scale doesn’t necessarily mean less enthusiasm. In fact, it indicates that by taking the right transformational steps to drive GenAI adoption now, the ROI will follow.

To learn more about GenAI right now, check out the State of Generative AI in the Enterprise survey or listen to our Deloitte TECHTalks series on GenAI. To explore Deloitte’s GenAI capabilities, visit our Generative AI page on Deloitte.com.