2023 will be remembered as the year when generative AI went mainstream. Inspired by successful cases of ChatGPT, companies began adopting, launching, and applying generative AI to enterprise applications. As we enter 2024, companies are now seeking to fully realize the promises of generative AI by integrating it more into their workflows.

However, a survey conducted by Forrester Consulting with 220 AI decision-makers in North American companies found that many companies still have concerns about the risks associated with this technology and see barriers to its adoption.

The survey highlights the main obstacles, including well-known issues such as biases, which keep organizations in the exploration or experimentation phase, preventing them from truly deploying foundational models into planned use cases.

If teams plan to double down on generative AI, these are the challenges they will have to address.

Understanding the transformative potential of generative AI

Given the abundance of successful cases on the internet, organizations across industries have already understood the transformative potential of generative AI.

In a Forrester survey conducted on behalf of Dataiku, 83% of respondents stated that they are exploring or experimenting with generative AI.

At the same time, over 60% of respondents claimed that they consider generative AI to be crucial or extremely important to their business strategies and plan to increase investments in data/AI initiatives by up to 10% in the next 12 months.

Leaders also emphasized that they already have use cases in progress. Over half of the respondents stated that they have identified multiple potential applications for this technology, including enhancing customer experience (64%), product development (59%), self-service data analysis (58%), and knowledge management (56%).

"This reflects an exploratory and curious sentiment, where organizations are attracted by the breadth of potential applications and expect to fully embrace the diversity of its transformative capabilities in the next two years," the survey pointed out. Respondents also mentioned that the expected broader benefits of these applications include enhancing existing products/services, creating new products/services, and optimizing internal and external operations.

Barriers to adoption still exist

Despite the optimistic outlook, leaders pointed out some barriers on the path to successful adoption of generative AI, including risks of violating data protection and privacy laws (31%) and challenges in developing skills and managing the complexity of generative AI (31%).

Over 50% of respondents also emphasized the risks of biases and illusions affecting the quality of generative AI outputs.

More importantly, when organizations fail to provide the necessary infrastructure for adopting generative AI, all these risks are further amplified. The lack of robust data infrastructure was identified as the biggest obstacle in this area mentioned in the survey.

Up to 35% of respondents listed insufficient infrastructure to support the consumption, storage, and sharing of large amounts of data as a pain point.

An equal number of respondents also mentioned difficulties in integrating with existing infrastructure, while 27% of respondents pointed out limitations in computing power.

Other barriers they mentioned involve handling governance mechanisms (35%), explainability and interpretability of AI (25%), talent and skill gaps (31%), and scalability of models.

"Organizations can mitigate many implementation challenges by adopting an approach that provides a collaborative set of functionalities. They can achieve this with AI platforms that offer pre-packaged solutions for accelerated development, structured environments for seamless integration, and robust frameworks and security features for standardization, governance, and compliance," the survey pointed out.

According to McKinsey, generative AI alone could increase global enterprise profits by $26 trillion to $44 trillion annually. It is estimated that this technology may have the greatest impact in the banking, high-tech, and life sciences sectors.