A survey of 300 IT leaders finds 90% are leading teams that use artificial intelligence (AI) tools to develop applications, with 71% reporting they will not consider hiring application developers who lack AI and machine learning experience.

Conducted by the market research firm Dynata on behalf of App Builder, a unit of Infragistics, the survey also finds 45% of survey respondents who lead organizations that don’t currently use AI to build software expect they will be within the year.

Overall, the survey finds 30% of respondents reporting that recruiting qualified developers and IT staff is among their top challenges in 2025. Beyond AI and machine skills, cloud computing (53%), problem solving (35%) and secure coding practices (35%) are among the top developer skills companies sought, according to the survey.

Nearly every IT leader of an organization uses AI in application development to improve security, with assessments (64%), testing code (60%), detecting breach patterns (59%) and scanning code for vulnerabilities (58%) being the top use cases.

Other use cases include automating mundane and repetitive tasks (40%), creating layouts and pages (34%) and detecting bugs (32%).

However, less than a third of respondents (30%) said automation is freeing up developers to focus on more strategic work.

Infragistics CEO Dean Guida said that, in general, it’s clear AI is transforming application development, but using it to reduce the number of application developers on staff is likely to prove counterproductive. In fact, organizations will need more developers who understand how to solve problems by applying critical thinking to how code is built than ever, he added.

The simple truth is that most organizations have a backlog of application development projects that are not going to be addressed solely using AI, said Guida. Application developers and software engineers will still need to review everything before deploying software in a production environment, he added.

Unfortunately, advances in AI may be turning students away from studying computer science in general and coding in particular, said Guida.

Hopefully, smaller organizations that would never have attempted to build custom applications will now do so using AI tools. The challenge they will encounter is that building and deploying software requires more than using an AI tool to generate code. Much of the benefit of AI will come from automating tedious tasks across the entire software development lifecycle, noted Guida.

It is still early days so far as the adoption of AI tools and platforms to build software is concerned, but it’s clear the proverbial genie is not going back in the bottle. The challenge and the opportunity now is to embrace AI to better understand its benefits and potential drawbacks. After all, with each additional advance, AI tools and platforms are only going to become able to automate a wider range of tasks. The issue then becomes not so much performing every task but rather making sure the output generated by these tools achieves the level of quality required. Otherwise, the initial investment made in AI tools will never actually be realized, no matter how much code is generated.