Paradox: Agentic AI dev roles are less in demand as agents take over

IEEE survey of senior techies in six countries finds recrutiment for data analytics, and machine learning on the up

Demand for software development skills in AI-related roles is set to fall next year as agentic AI accelerates across business markets, according to an IEEE industry survey.

Or so says research by the global technical professional organization, which polled 400 CIOs, CTOs and IT directors in Brazil, China, Japan, India, the UK, and the US.

IEEE states that nearly all professionals expect agentic AI innovation to continue at "lightning speed" in 2026, as both established enterprises and startups deepen investments and commitments to the technology.

The technology is reshaping hiring priorities in tech departments. Demand for software development skills in AI roles dropped 8 percentage points to 32 percent compared to last year. Meanwhile, employers increasingly seek AI ethics expertise (44 percent, up 9 points); data analysis skills (38 percent, up 4 points); and machine learning capabilities (up 6 points).

Paradoxically, 39 percent of respondents plan to use agentic AI to aid software development itself — suggesting the technology may be replacing the very skills it requires. Only real-time cybersecurity vulnerability identification and attack prevention ranks higher at 47 percent, although this figure has fallen.

The sphere of AI may be creating demand for skills in certain areas, but in terms of software development its impact seems to be mixed, at best, while other doubt its value. Oracle's CTO and co-founder Larry Ellison has claimed most of the company's new code is now AI-assisted, while Salesforce has launched Agentforce Vibes, by which it means AI-assisted, rather than an attempt to encourage prompt-to-production development practices.

The IEEE report says the top industries that expect to experience the most significant transformation from AI next year will be software (cited by 52 percent); banking and financial services (42 percent); healthcare (37 percent) and automotive and transportation (32 percent).

Elsewhere in the report, tech leaders see an appetite for humanoid robots. Seventy-seven percent of respondents agreed "the novelty of humanoid robots can inject fun into the workplace" but over time they would "become like commonplace co-workers with circuits."

Back in June Garter estimated that two in five agentic AI projects would be ditched by the end of 2027 due to rising cost, uncertain commercial value or insufficient risk controls. And it found AI agents get office tasks wrong around 70 percent of the time.

Last week, Forrester said it expects a number of businesses to "quietly rehire" employees that were given the elbow in the name of AI efficiency. ®