While Generative AI Simplifies Tasks, Challenges Abound For Developers: IBM Survey

As enterprises move to adopt generative artificial intelligence (AI), they are assigning their developers the responsibility of building, customising, testing, and deploying generative AI applications. However, businesses underestimate the complexity of the AI development process and stack. 

A new survey of enterprise AI developers sponsored by IBM explores that complexity and shows the challenges developers are facing when it comes to skills variance, complicated toolsets, and ensuring accurate and trusted results. 

Intriguingly, according to the survey, a majority of developers who identify as “AI developers” or “data scientists” view themselves as experts in generative AI. On the other hand, app developers rarely view themselves as generative AI experts, despite being on the front lines of generative AI adoption. Only 24% of app developers surveyed ranked themselves as “experts” in generative AI.

This illustrates the skills gap in generative AI. This is unfamiliar territory for developers, and there is a steep learning curve. 

The lack of clarity about trustworthy frameworks and toolkits exacerbates the skills gap. Lack of a standardised AI development process and developing an ethical and trusted AI lifecycle that ensures transparency and traceability of data were top challenges in the development of generative AI applications for 33% of those surveyed.

The available tools are a source of frustration for developers. Performance (42%), flexibility (41%), ease of use (40%), and integration (36%) were the four most essential qualities in enterprise AI development tools, according to those surveyed. Yet over a third also said these qualities are the rarest.

A majority (72%) of those surveyed use between five and 15 tools to create an AI enterprise application. A notable number—13%—use 15 or more tools. This signifies that developers face complex challenges in the AI stack, which drains enterprise investment in AI.

As the industry moves closer to agentic AI, which offers more autonomy and power but also depends on trust and integration with IT systems, these challenges will become worse, the survey noted.

Almost all developers surveyed (99%) are exploring or developing AI agents, and the top concern reported for agentic development is trustworthiness.

Developers want tools that are simple to learn, especially considering how quickly the generative AI ecosystem is changing. Only one third of those surveyed are willing to invest more than two hours in learning a new AI development tool. This indicates that, for wider use, simplicity and user experience are important when introducing new technologies to support the AI development process.

Regarding developer productivity, the survey found widespread adoption and significant time savings from the use of AI-powered coding tools. Of those surveyed, 99% use coding assistants in some way for AI development. And they have emerged as a huge benefit. These tools saved them 1-2 hours per day (41% of developers), with some saying it saves them 3 hours or more (22%).