We know artificial intelligence – particularly generative and agentic AI – is reshaping jobs. But the exact impact is still a great unknown. But the impacts we’re already seeing in tech jobs, many of which are at the forefront of the AI, generative AI, and agentic revolution – may provide clues to where things are going – a crucible for the AI-shaped job market of the near future.
For starters, there doesn’t appear to be evidence that AI is sweeping away jobs. There’s even some evidence that it may help increase, rather than reduce jobs, particularly for technology occupations. There has been no noticeable impact on graduates starting out in the job market, and there’s even been growth in white-collar jobs, an analysis published in The Economist shows. The researchers cite the relative immaturity of AI development – only 10% use AI on a enterprise scale – and it’s primary role as a productivity platform.
In addition, looking more closely at tech roles, at least seven in 10 technology leaders surveyed by one major analyst firm, 69%, indicate they’re planning to increase headcounts – at least within technology areas – to build genAI capabilities.
Technology jobs are the first category being reshaped by AI, a recent study out of the Federal Reserve Bank of Atlanta confirmed. The category of “computer and mathematical occupations” saw demand for AI skills grow from two percent of postings in 2010 to 12% of postings in 2024.
Other occupational groups, including architecture and engineering; business and financial operations; and management, are also seeing increasing proportions of AI within their job descriptions.
AI and genAI “are already changing the set of skills employers are demanding from the workforce,” the Fed survey suggested, with the percentage of job postings requiring AI-related skills increasing steadily. “Demand for AI skills is rising not just in computer and mathematical occupations but in a broader set of occupations, which they attribute to the increasing technical capabilities of AI to perform more tasks.”
Industry observers point to technology roles as examples of how jobs are evolving to hybrid mixes of human and genAI and agentic AI-led tasks.
Notably, the latest evolution of AI – agents – are poised to take on more tasks within a range of jobs. AI agents “can take a goal, break it into subtasks, and work on finding the best solution for these tasks individually,” aid Andreas Welsch, founder and chief AI strategist at Intelligence Briefing. “Agents have access to additional information, tools, and resources – for example, code repositories, APIs, or websites. They can take on specialized roles such as an architect, software engineer or QA tester, and work on tasks within the typical scope of that role.”
This doesn’t mean AI will pick up tasks and business will go on as usual. “Firstly, it is a complete paradigm change in how we use and interact with software systems,” said Chris Burchett, senior vice president for genAI at Blue Yonder. “Secondly, it is evolving at an unprecedented pace never before seen.”
To break in and thrive in such a world, Burchett advises “not to wait. You have to get started using the technology immediately. Second, you must have staying power to evolve with the changes because that is the only way to keep up and learn the unique capabilities AI unlocks. Third, you need to an abstraction layer that allows you technical agility to move across different models, frameworks and providers with minimal rework.”
At the same time, the role of AI has limits. “AI might initially perform at a junior coder’s level, “but still requires human input and oversight,” Welsch pointed out. “This means that human software developers will still need to define the project, its objectives and personas, and the expected behavior of an application. Users will need to acquire this knowledge as well as learn how to communicate with agentic AI systems to derive the most relevant results quickly.”
While large language models have been trained on historic data and are able to generate code, “this code is not always the most efficient implementation of a solution,” he added. Importantly, “just because the AI-generated code is functional doesn’t automatically make it secure. Additional tools or humans in the loop are needed to conduct security reviews of the generated code to mitigate any loopholes.”
AI in general, “has the opportunity to amplify – not eliminate – human talent,” said Gajen Kandiah, AI and enterprise transformation leader and former president and COO of Hitachi Digital. “This is not about whether AI replaces developers. It is about how the role of developers—and the systems they create—are being redefined. The truth, as with most meaningful shifts, sits in the nuance. We will not see the future of engineers vanish. Instead, they’ll evolve into AI trainers, strategic integrators, and problem-solvers.”
One thing is clear, Kandiah continued. “The best developers will not be those who write the most lines of code – but those who design and deliver the most impact by partnering with intelligent systems.” This applies to all workers as well, as the ability to work with AI to create new approaches to problems and opportunities will be a necessity in the months and years ahead.