How to keep your IT talent pipeline from collapsing

The transformative lure of AI is rapidly pushing IT leaders’ talent pipelines toward more of a crossroads than many may fully want to admit.

The traditional approach of growing IT expertise in-house from entry-level positions is being challenged by a combination of skills-demand shifts toward AI experience and the replacement of entry-level roles in favor of AI automation.

Employment among early-career workers, ages 22 to 25, in the most AI-exposed occupations has fallen 16% since the introduction of ChatGPT in late 2022, according to a widely cited study from Stanford’s Digital Economy Lab. For entry-level software developers, the drop was nearly 20%. As the pool of talent with early-career IT pros with hands-on experience shrinks, IT leaders are likely to face stiffer challenges filling more vital midlevel roles down the road.

Looking forward, some IT leaders believe replacing junior engineers and other entry-level IT roles with AI to cut costs will eventually backfire, leaving companies short of experienced staff who can tackle difficult problems and design scalable solutions.

According to a recent Gartner survey of global business executives, organizations that automated aspects of their businesses and reduced their workforces aren’t seeing returns from those supposed efficiencies. What has improved the bottom line? Investing in new roles, upskilling, and systems that amplify the capabilities of staff so they can supervise and grow autonomous work.

Moreover, the Gartner report forecasts that autonomous business practices will require more staff, not less, over the next two to three years, leading to a net positive in job growth as people are hired to manage those efforts.

Yet, in the short term, investors are rewarding companies that make AI-related workforce reductions. And many executives are pushing for the same. So how are CIOs and other leaders planning to build the necessary skills for future success by creating a pathway for middle- and senior-level IT talent?

‘Early in context’

In response to this downward trend in early career hiring, Microsoft’s Mark Russinovich and Scott Hanselman penned an article that pushes back on this trend. They propose bringing in early-career programming talent and pairing them with experienced mentors on product teams, where they can help new hires identify — and solve — real-world problems that AI might miss.

In the article, the Microsoft execs noted that experienced programmers found dozens of problems in AI-generated code that appeared to work correctly. They also pointed to the risk of “cognitive debt,” citing MIT research that found reduced brain activity among people relying heavily on AI for writing tasks.

“While agents can speed up workflows and reduce manual effort, they lack the intuition to anticipate edge cases and build robust solutions,” the authors wrote. “Relying too much on AI risks missing subtle bugs, architectural flaws, and vulnerabilities that only skilled engineers can catch. Human oversight, critical thinking, and domain knowledge are indispensable for both correcting errors and driving innovation as technology progresses.”

Hanselman, vice president and member of technical staff at Microsoft, argues that software development isn’t simply a matter of writing code. Senior engineers, he notes, have experience in what works, what fails, what can break in production, and what elegant design looks like — and how to scale it. AI can increase output, but it does not help a new developer learn this sort of judgment.

“When you say early in career, it’s actually early in context — junior devs are missing context,” he says. “The way that we develop good taste is through failing in a safe place. And right now, companies hire juniors, throw them at a problem, chew them up and spit them out — and that’s the wrong way to do it.”

A new mentorship model

Hanselman suggests, instead of slashing roles for junior programmers, companies should be creating systems that help them develop the skills necessary to become valued senior contributors in the future.

He proposes adopting a mentorship approach called a “preceptorship,” borrowed from the medical field, where senior engineers are explicitly responsible for helping juniors gain experience and develop good judgment. Hanselman’s wife is a nurse and preceptor, and her experience helped spur the idea.

“The preceptorship acknowledges that a nurse has passed the board,” he explains. “They’ve joined the company. It’s their first day on the job. They are qualified to be there. They are supposed to be there — but they’re missing context.”

Technology companies need a similar model, he argues, where programmers are allowed to learn, not just produce, from experienced mentors: “We need high communicators, with high agency — kind individuals who will invest in the future.”

He contrasts this practical, real-world mentoring approach with a coding boot camp.

“What do people do in boot camps? They wash out,” he says. “You couldn’t hack it. A preceptorship is a relationship between a senior engineer, who has your best interest at heart and is going to help you become a better AI-augmented software engineer — not a vibe coder. We’re not vibing into production. We are using the powerful tools that have been developed to create high-quality software with good taste and with good discernment at scale.”

A talent gap in the making

Companies that eliminate junior roles because AI can do some entry-level tasks may see improved short-term output while weakening their future technical capabilities. Tech executives say a lack of investment in early career hiring will show up in the future as a dearth of leadership and institutional knowledge, as well as a reduction in product quality and the ability to effectively manage and oversee code or other work created with AI.

“Senior engineers are built through exposure to real systems, not just writing code,” says Craig Miller, former CIO of fast-food chain Sonic, now a consultant, board advisor, and author. “They need to understand how things scale, how they break, and how decisions impact the business. That experience cannot be automated.”

Reducing junior developer roles should be seen as a long-term capability risk instead of a budget efficiency, says Macaire Montini, vice president of people and culture at cloud-based HR software company HiBob.

“The decline in junior developer roles isn’t just an employment trend,” Montini says. “It’s a long-term pipeline problem that technology leaders should treat with the same urgency as any infrastructure risk. If you stop bringing in early-career talent, you don’t just have a gap today — you have a leadership drought in five years.”

Zsolt Kerecsen, CTO at Graphisoft, argues that replacing early-career staff with AI hurts staff growth and undercuts an organization’s ability to manage autonomous capabilities. CIOs should treat early-career hiring as an investment in future delivery quality, system oversight, and AI governance, he says.

“Experienced developers are needed to train AI and validate its outputs,” he says. “That’s why trying to substitute juniors with AI is a fundamentally flawed approach. Instead, AI should be used — guided by seniors — to support junior developers and help them become seniors more quickly.”

Miller says the reduction in early career hiring is just one sign of a broader issue of “slow decay,” where current tech staff aren’t training their replacements. He points to other indications of a future talent crisis: “Decline in CS enrollments as prospective students respond to deteriorating job market signals, which could produce a senior engineer shortage in 5 to 10 years even as AI reduces demand for entry-level workers today. The real risk is not that AI will eliminate the need for developers. It’s that companies will eliminate the early learning ground that has always produced great ones.”

Filling the pipeline

With early-career roles evolving quickly, experts advise CIOs to take a more intentional approach to hiring and training IT talent, programmers in particular — one that uses AI to help junior staff become better, faster, instead of replacing them.

AI may enable junior developers to take on more advanced tasks earlier, Montini says, but they still need mentoring and structured guidance to become experienced contributors.

“We believe the answer isn’t just hiring,” Montini says. “It’s how you onboard and develop early-career talent once they’re through the door. Structured training, clear skill development pathways, and meaningful mentorship are what actually close the gap between potential and performance. Without that scaffolding, junior hires churn before they become the midlevel talent you need.”

Paul DeMott, CTO at Helium SEO, says organizations should rethink talent development from a new hire’s first day.

“Before a junior developer on our team writes a single line of code on any new feature, they have to propose the full architecture for it, present it in a 15-minute review with the senior team, and explain every tradeoff they considered,” he says. “The junior does not implement anything until they defend those decisions. This process forces systems thinking before syntax thinking, which is exactly what separates a developer who grows into senior roles from one who stays at the execution layer indefinitely.”

In the past year and a half, DeMott says, that process has helped junior hires rise more quickly through the ranks, with two junior developers promoted to midlevel roles.

Kerecsen says his company actively seeks out junior talent at the university level, works with them for several years, then brings them on as junior or potentially midlevel engineers.

“There is a concerning misunderstanding about AI’s potential, especially regarding its ability to replace junior developers,” Kerecsen says. “It is actually disastrous for delivery quality and long-term sustainability. Junior developers are an investment in our future.”

Liz Eversoll, CEO of upskilling and recruitment company Career Highways, says organizations should move from informal apprenticeship to a more intentional model for skills-based growth.

“The next generation of senior programmers will be developed differently,” Eversoll says. “Junior engineers can now contribute to higher-complexity work earlier by using AI as a copilot, but that only works if organizations provide pathways that connect real work, learning, and continuous assessment.”

Building judgment, not just output

The goal is to help junior developers gain the kind of experience that allows them to understand systems, weigh tradeoffs, and eventually guide technical decisions.

Former Sonic CIO Miller says that kind of experience cannot be automated.

“The organizations that get this right will balance AI-driven efficiency with structured mentorship and real-world exposure, treating talent development as a long-term priority,” Miller says. “The next generation of senior engineers will not emerge accidentally. They will have to be built through structured apprenticeship, guided use of AI, real exposure to production environments, and deliberate development of judgment, architecture thinking, debugging discipline, and business context.”

Rema Lolas, founder of team-building platform Groziac, says AI may make technical skills more accessible, but it will also put more pressure on how people work together.

“AI may level the technical playing field, but it will amplify the differences in human performance,” Lolas says. “The organizations that recognize this early will stop treating development as a training problem, and start treating it as a system design challenge — where people are intentionally developed not just in skill, but in how they operate and perform together.”

Microsoft’s Hanselman says the skills that matter most today are not just AI prompt fluency or the ability to generate code quickly, but systems thinking and communication.

“So for the young person who’s coming into this, you can’t have blinders on,” he says. “Making large, interesting systems that help people and make their lives better — that is not being commoditized. You need big-picture thinking, taste, discernment, good judgment, good communication skills, and a rock-solid understanding of the basics. Just because I’m riding around in an Uber doesn’t mean that I don’t know how to change a tire.”

Tech leaders say organizations need to make early-career growth a core part of engineering work. That means giving junior staff real programming work, in-the-moment senior guidance and AI support that accelerates learning without replacing it.

“Ultimately, developing senior talent is no longer a byproduct of hiring, it’s the result of deliberate infrastructure,” Eversoll says. “Organizations that invest in skills-based progression systems will not only sustain their pipeline, but accelerate it.”