The question facing employers in 2026 is becoming less about whether artificial intelligence can perform a task and more about where humans still need to remain in control. New data from recruitment technology company Xobin suggests that distinction is becoming critical: 72% of skill groups in its 2024 hiring framework were fully or partially delegable to AI under at least one tested model.
The finding comes from Xobin’s Human vs AI Skills Report: 2026 Mid-Year Edition, based on data through June 30, 2026. The study examined 683 skill groups, 113 leadership scorecards from 92 employers and thousands of technical assessment requests.
AI is changing the skills employers look for, but the more consequential shift may be happening inside the hiring process itself.
Xobin’s latest research points to a labor market where the boundary between human work and machine-assisted work is becoming increasingly difficult to define. Rather than suggesting that AI is simply replacing occupations, the report examines whether individual skills can be delegated to AI and where human judgment remains necessary.
That distinction matters.
A task being technically delegable does not necessarily mean an organization can remove the person performing it. Xobin notes that partial delegation can still require someone to frame the problem, evaluate an AI-generated result and intervene when the system produces an unsuitable answer.
The degree of delegability varies considerably across skill categories.
Analytical reasoning showed the highest levels, with 90% to 100% of the skills in the category considered fully or partially delegable under at least one tested AI model. Operations and execution were considerably less exposed, at 28%, while leadership-related skills remained below 50%.
The numbers offer a different way to think about AI disruption.
Instead of asking whether a job is “AI-proof,” employers may need to determine which components of a job require independent human capability and which can be augmented or delegated to AI.
That could change how organizations write job descriptions, structure interviews and design technical assessments.
Hiring for AI judgment, not just AI usage
The shift is already visible in coding assessments.
Xobin found that traditional AI-free coding tests accounted for between 75% and 100% of technical assessment requests during the first half of 2024. By the first half of 2026, that share had fallen to 25% to 50%.
At the same time, AI-assisted coding tasks — including generating, testing, debugging and explaining code — rose from less than 25% of requests to between 50% and 75%.
That suggests employers are beginning to test something different.
The ability to write code without assistance remains relevant, but organizations increasingly need to know whether candidates can work effectively alongside AI. That means evaluating how applicants prompt systems, inspect generated code, identify errors and make decisions when an AI tool produces an unreliable answer.
For recruiters and talent leaders, this creates a new assessment challenge: how do you distinguish someone who can use AI productively from someone who simply knows how to ask an AI tool for an answer?
The answer increasingly involves assessing judgment.
Human skills remain part of the leadership equation
The report’s leadership findings reinforce that point.
Across 113 leadership scorecards from 92 employers, emotional intelligence-related criteria represented an average 52% of scorecard weighting, narrowly exceeding all other criteria combined.
That puts interpersonal capability directly inside the formal definition of leadership fit used by the employers represented in the dataset.
It also aligns with a broader shift toward evaluating skills that are difficult to reduce to automated execution.
Collaboration, for example, appeared across 12 to 17 of 24 tracked emerging skill groups in 2026, compared with four in 2024. Non-linear thinking increased from zero to five groups in 2024 to six to 11 in 2026.
These changes suggest that as AI takes on more predictable analytical and technical work, organizations may place greater emphasis on skills associated with coordination, interpretation and complex decision-making.
That does not make those skills uniquely human. AI systems can increasingly simulate communication and reasoning. But organizations still need people who can establish context, decide what matters and take responsibility for outcomes.
The rise of the translator
Another notable finding is the growing importance of business-to-technical translation.
Among 35 technical roles analyzed by Xobin, the skill appeared in 10 to 19 roles, putting it among the top five requested capabilities. It remained behind AI integration and automation and analytical problem-solving, which appeared in 30 to 35 roles.
The trend reflects a growing requirement for technical professionals to understand business objectives rather than operate solely within technical boundaries.
As AI makes software development and other technical workflows faster, the bottleneck can move elsewhere. Someone still has to translate a business problem into a technical requirement, determine whether an AI-generated solution addresses the actual problem and explain the trade-offs to stakeholders.
That makes cross-functional fluency increasingly valuable.
Implications for HR technology
For HRTech vendors, the findings could accelerate the evolution of talent assessment platforms.
Traditional assessments often measure whether a candidate can independently complete a defined task. But in AI-enabled workplaces, employers may need multiple assessment modes: independent performance, AI-assisted performance and the candidate’s ability to evaluate AI output.
That creates opportunities for recruitment and assessment platforms to measure AI literacy, verification skills, critical thinking and human-machine collaboration alongside conventional technical competencies.
It also raises an important governance issue.
If AI is involved in assessing candidates, employers must distinguish between evaluating a candidate’s ability to use AI and allowing AI to make the hiring decision. Transparency, consistency, explainability and human oversight remain important as automated tools become more deeply embedded in talent acquisition.
Xobin’s research ultimately points toward a more nuanced model of workforce planning. AI does not need to eliminate an entire job to fundamentally change it. Delegating 30%, 50% or even 70% of a workflow can alter the skills required, the number of people needed and the way performance is measured.
For employers, the practical question is therefore becoming more specific: Which work should AI perform, which work should humans perform independently, and where should humans supervise the machine?
That distinction could become one of the defining principles of hiring in the AI-first workplace.
Market Landscape
Xobin’s findings fit into a wider transition in talent acquisition from job-based hiring toward skills-based workforce planning.
The major HCM and talent platforms — including Workday, SAP, Oracle, Eightfold AI and LinkedIn — are increasingly incorporating AI into recruiting, skills intelligence, talent matching and workforce planning. The competitive question is shifting from simply automating recruiting administration to understanding how AI changes the underlying composition of work.
Three developments stand out:
- Assessment is becoming AI-aware: Employers can no longer assume candidates will complete technical assessments without AI assistance.
- Human oversight is becoming a skill: Verification, judgment and exception handling are increasingly important complements to AI-generated output.
- Skills taxonomies are becoming dynamic: Job descriptions built around static responsibilities may become less useful as AI changes task composition.
The most important caveat is that Xobin’s 72% figure measures skill-group delegability, not jobs eliminated or employees replaced. It should therefore be interpreted as an indicator of task-level AI exposure rather than an employment forecast.
Top Insights
- Xobin found 72% of assessed skill groups could be fully or partially delegated to AI under at least one tested model, highlighting rapid task-level transformation.
- AI-free coding assessments are declining as employers increasingly test candidates on generating, debugging, testing and explaining AI-assisted code.
- Emotional intelligence represents 52% of average leadership-scorecard weighting, showing interpersonal capability remains central to hiring decisions.
- Business-to-technical translation is becoming a top-five skill across studied technical roles as companies seek stronger links between technology and business outcomes.
- Collaboration and non-linear thinking are appearing more frequently in emerging roles, suggesting job design is adapting alongside AI adoption.
Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights
