Generative AI is reshaping industries, but IBM shows why it’s harder to build than you think

Generative AI has emerged as a transformative force in enterprise innovation, revolutionizing industries with applications that streamline workflows, improve decision-making, and enhance customer experiences. However, beneath the sleek interfaces of these cutting-edge technologies lies a complex development ecosystem that is far from simple. A recent survey sponsored by IBM and conducted by Morning Consult offers an in-depth look at the challenges enterprise developers face when building generative AI systems.

The study, which included responses from over 1,000 U.S.-based enterprise AI developers, highlights significant issues, including skills gaps, tool inefficiencies, and trust concerns. These challenges raise critical questions about the sustainability of generative AI development as businesses increasingly depend on the technology for competitive advantage.

Skills Gaps: The Steep Learning Curve of Generative AI

The survey reveals a wide disparity in generative AI expertise among developers. While AI developers and data scientists often identify as experts in the field, many application developers—the professionals on the frontlines of implementation—report limited familiarity with the technology. Only 24% of application developers surveyed described themselves as experts in generative AI.

This skills gap presents a substantial barrier, particularly as generative AI requires advanced technical knowledge, including understanding machine learning algorithms, data architecture, and ethical considerations. The rapid pace of innovation further exacerbates the issue, leaving developers scrambling to keep up with new tools and frameworks.

According to the survey, developers are calling for standardized development processes to mitigate these challenges. The lack of clear guidelines and reliable frameworks not only hinders learning but also complicates efforts to ensure transparency and traceability, two critical factors in building trusted AI systems.

Tool Complexity: The Hidden Burden on Developers

Developers also face significant challenges with the tools they use to build generative AI systems. While qualities such as performance, flexibility, ease of use, and seamless integration are considered essential for enterprise AI tools, these features are often difficult to find. The fragmented nature of the AI development ecosystem forces developers to juggle an average of five to 15 tools per project, with some relying on 15 or more.

This reliance on numerous tools creates inefficiencies and complicates workflows, slowing the development lifecycle. The survey found that while developers prioritize tool simplicity, only a third of respondents are willing to spend more than two hours learning a new AI development tool. This finding underscores the urgent need for intuitive, user-friendly solutions that allow developers to focus on innovation rather than navigating complex toolsets.

Trust and Transparency: The Cornerstones of Agentic AI

Trustworthiness emerged as a top priority for developers working on agentic AI systems, which are designed to operate autonomously. Nearly all respondents (99%) reported involvement in developing AI agents, but many cited concerns about ensuring these systems integrate effectively with broader IT environments.

Trust is a cornerstone of successful AI adoption, especially as enterprises leverage agentic AI for critical operations. Developers must navigate not only technical challenges but also ethical considerations, such as ensuring data transparency and avoiding biases in AI-generated outputs. Addressing these concerns requires both robust frameworks and reliable tools that prioritize ethical AI practices.

Simplifying the AI Stack: IBM’s Approach

IBM has positioned itself at the forefront of efforts to simplify generative AI development. Through its watsonx.ai platform, IBM provides developers with an end-to-end suite of tools, frameworks, and integrations that streamline the AI development lifecycle. The platform emphasizes open-source technologies, such as its Granite models, which promote transparency and collaboration.

IBM’s watsonx Code Assistant further enhances productivity by automating aspects of coding, allowing developers to focus on higher-value tasks. Meanwhile, IBM’s Application Integration solutions offer seamless API management, enabling developers to modernize their AI applications across hybrid environments.

The survey also highlights the benefits of AI-powered coding tools, with 99% of respondents reporting their use in some capacity. These tools save developers significant time, with 41% of respondents estimating daily savings of 1–2 hours and 22% saving 3 hours or more.

The Future of Generative AI Development

As generative AI continues to evolve, simplifying the development lifecycle is crucial for unlocking its full potential. By addressing skills gaps, reducing tool complexity, and prioritizing trust and transparency, enterprises can empower their developers to innovate more effectively.

IBM’s focus on open-source AI and user-friendly development tools reflects a broader industry shift toward creating sustainable, ethical, and efficient AI systems. Simplifying the AI stack is not just a technical necessity but also a strategic imperative for enterprises looking to harness the transformative power of generative AI.