Software engineering roles shift toward customer-facing AI oversight
AI-generated coding is pushing software engineers away from solitary programming and toward customer conversations, product judgment, and oversight of automated tools. The shift is changing hiring criteria, office-work expectations, and career paths across the software sector, even as overall demand for developers remains strong.
Highlights
- AI tools are reducing time spent on manual coding, with engineers spending more effort reviewing AI output, communicating with customers, and informing product decisions.
- Job postings for in-office tech roles climbed from 63% in early 2024 to 84% for fall, while remote postings fell from 10% in early 2023 to 3%, reflecting a hiring shift.
- Anthropic's junior technical hires dropped from 27% in March 2025 to 9% in May as automation narrows entry-level opportunities and increases demand for communication skills.
AI tools reshape engineering work
As reported by Business Insider, developers are increasingly spending less time writing code line by line and more time reviewing AI output, explaining technical decisions, and engaging with users and customers.Michael Irwin, a principal software engineer at Docker, recently used a handmade gaming controller at a conference booth to draw attendees into conversations about AI agents, illustrating how presentation and interpersonal skills are becoming more relevant in technical roles. He says AI has made side projects like that easier to build, while shifting more of the detailed coding work to automated systems.
Engineers interviewed in the report say the job now requires deeper focus on why products are being built, not just how to code them. Paul Elliott, vice president of product and engineering at Chromatic, says developers need broader business understanding and must ask better questions across teams. Marc Brooker, vice president and distinguished engineer at Amazon Web Services, says engineers on his team now spend more time with customers and prospects, then bring those insights back into product and technical decisions.
The change is also blurring traditional job boundaries. Dylan Wight, founding software engineer at Flagler Health, says his work now includes reviewing issues raised by agents and meeting regularly with healthcare users. Frank Fusco, CEO of Silicon Society, says engineers are taking on more product autonomy while product managers are doing more technical work, reflecting shifting definitions for both roles.
Hiring patterns and career risks emerge
Employers are increasingly screening for communication ability alongside technical expertise, while also pulling engineering roles back into offices. Michelle Volberg, CEO of recruiting firm Twill, says about half of the roles her company filled last year were remote, but roughly 95% of openings now require hybrid or full-time in-office work.Data cited in the report points to the same direction. Robert Half says 63% of tech job postings had in-office requirements in early 2024, rising to 84% this fall, while the share of remote tech postings fell from 10% in early 2023 to 3%. CompTIA data also shows communication skills appeared in 47% of open software engineer job descriptions in early 2020, compared with nearly 60% now.
At the same time, the article says opportunities for junior engineers are narrowing as companies use AI to accelerate coding and rely more heavily on experienced staff to review machine-generated output. Anthropic's workforce report showed technical workers spending 16% more time in meetings than a year earlier, while the proportion of junior technical hires there fell from 27% in March 2025 to 9% in May.
Despite concerns about automation, software employment is still expanding. The U.S. Bureau of Labor Statistics says about 1.9 million people worked in software development, testing, and quality assurance in the U.S. as of last year, up from about 1.5 million in 2020, and projects 10% industry growth over the next decade. The longer-term risk described in the report is not only displacement, but a weaker training pipeline for future engineers who need enough practical experience to catch flaws in AI-generated code.
We have also analyzed Gartner’s AI job gains forecast for 2025 to 2030, which highlighted how roles and required skills may shift.