The human + AI workforce transformation is here. Is your hiring process ready?

Download the full 2025–2026 AI Workforce Transformation Report →
https://karat.com/resource/ai-workforce-transformation-report/

AI is transforming software development faster than most organizations can adapt. Engineering productivity is accelerating, the ROI of strong engineers is increasing, and AI-native workflows are reshaping how teams build software. But, while engineering has moved into the human + AI era, most hiring practices have not.

Our new AI Workforce Transformation Report, based on insights from 400 engineering leaders across the U.S., India, and China, highlights an urgent disconnect: AI is magnifying the impact of top engineers. Yet, most organizations still don’t have a reliable way to identify AI-ready talent.

Below are some of the key findings and why they point toward a critical shift: the need for human + AI interviews that reflect the real-world workflows of modern engineering.

AI is boosting productivity (for engineers who know how to use it)

Across organizations, engineering leaders report an average 34% productivity lift tied directly to AI enablement. But the benefit is not evenly distributed. The strongest engineers: those who can integrate AI tools into design, debugging, refactoring, and decision-making, are pulling further ahead, while weaker and median engineers see far smaller gains.

As David Lau, VP of Engineering at OpenAI, explained:

“Frontier models are advancing so quickly that last month’s edge cases become this month’s baseline. We’ve moved from autocomplete to agents that write and refactor entire libraries, and we’re increasingly seeing models explore novel solutions to complex problems.”

AI isn’t replacing engineers. It’s amplifying them. But only if they know how to harness it.

Most companies still evaluate engineers as if it were 2019

Even as AI reshapes engineering, most hiring processes remain frozen in pre-LLM assumptions.

Our research found that:

• 60%+ of companies still prohibit AI use in technical interviews
• Only 27% of organizations are training interviewers to evaluate AI-enabled problem-solving
• Fewer than 30% are updating interview content to reflect AI-native workflows
  

As a result, organizations often miss the very talent they’re trying to hire.

They exclude engineers who work effectively with AI. They rely on puzzles and toy problems that modern models can now auto-solve. And they use automated evaluation systems that can’t distinguish between a strong engineer’s reasoning and a model’s autocomplete.

Asynchronous coding challenges are obsolete

One of the clearest signals in the research is the rapid decline in the value of static, asynchronous take-home projects and code tests.

David Lau put it bluntly:

“Asynchronous take-home coding challenges are losing their effectiveness as AI tools can now solve most of them with ease. Moving forward, the best way to evaluate a software engineer’s capabilities won’t come from grading homework, but from engaging in a high-bandwidth, real-time discussion about how they approach problems — and how they demonstrate pragmatism, efficiency, and resilience along the way.”

In an AI-native world, hiring must be dynamic, interactive, and adaptive, because that’s how engineering now works.

Why human + AI interviews provide the strongest signal

One of the clearest conclusions from the report is that the most predictive way to evaluate AI-ready engineers is to mirror real AI-native development environments.

That means humans and AI together, revealing engineering ability in context.

Here’s why this works:

1. AI-enabled interviews surface real collaboration skills

Top engineers treat AI like a teammate. Interviews that allow (and encourage) AI use show how candidates actually work and are more predictive of on-the-job performance.

2. Expert interviewers probe candidate judgment and reasoning

As models get stronger, the differentiator isn’t who can produce code; it’s who can make sound decisions. Human interviewers can probe trade-offs, question rationale, and evaluate adaptability in ways AI-only or fully automated tests cannot.

3. Multi-file, real-world scenarios resist model shortcuts

Static questions have become easy for models to brute-force. Complex projects with ambiguous requirements, architectural decisions, or debugging tasks reveal deeper engineering skills.

4. Human oversight ensures fairness as AI changes rapidly

As models evolve monthly, human-led interviews ensure interviews stay calibrated and unbiased.

According to Sagnik Nandy, CTO of DocuSign:

“AI is transforming engineering, but the real breakthroughs happen when human judgment and AI capabilities work together. What’s been missing is a way to measure that combination reliably.”

This is exactly what human + AI interviews are designed to solve.

Download the full report for more AI hiring trends

The full 2025–2026 AI Workforce Transformation Report dives deeper into:

• How AI is reshaping engineering productivity
• What separates AI-ready engineers from the rest
• Why strong engineers are now delivering triple the ROI
• Emerging hiring, upskilling, and organizational strategies
• What engineering leaders expect over the next 36 months

Read the full report →  https://karat.com/resource/ai-workforce-transformation-report/