The Announcement
A June 2026 study by GoHumanize, an AI humanizer tool, examined 60 professional skills to determine which are most resistant to automation. The study scored each skill across four dimensions: employer importance, job listing frequency, automation resistance, and human dependency. The top findings are pointed: leadership ranks as the hardest skill for AI to replace, with machines able to automate only 31% of what CEOs do. All ten of the most future-proof skills involve managing, communicating with, or understanding other people. Technical skills like data analysis, by contrast, rank near the bottom of the automation-resistance scale despite strong employer demand.
Our Analysis
This study arrives at a moment when enterprise AI adoption is accelerating at a pace that outstrips organizational readiness. The GoHumanize findings are worth taking seriously, but they deserve some analytical context before IT and business leaders decide how to interpret them.
The Study Has Real Limitations
GoHumanize is an AI humanizer tool, a product category that benefits commercially from a narrative in which AI-generated content needs human refinement. That doesn’t make the research wrong, but it makes independent corroboration worth seeking before treating the rankings as authoritative. The methodology relies on composite scoring across four variables, which means the rankings are sensitive to how those variables are weighted. The study does not disclose its weighting formula in the excerpts shared here.
The finding that data analysis is “one of the easiest skills for AI to automate” also deserves scrutiny. Large language models are good at pattern recognition and summarization. They are less good at the higher-order judgment required to determine which questions are worth asking of a dataset, how to account for data quality issues, or how to communicate findings to a skeptical executive audience.
What the Data Actually Tells ITDMs
The business-relevant insight here is not that leadership is safe and data skills are not. It’s that the economic value of human judgment is being repriced in real time, and most organizations are not accounting for that in their workforce planning.
ECI Research’s 2025 AI Builder Summit survey found that 35.8% of respondents strongly agreed that “this generation of business leaders will be the last to manage a workforce composed entirely of humans.” That’s a significant share of enterprise AI leaders who believe a structural shift in workforce composition is already in motion, not a distant scenario. Yet the same survey found that 44% of enterprise AI leaders have only moderate confidence that AI agents can act autonomously without human intervention. The implication is that enterprises are simultaneously planning for a mixed human-AI workforce and expressing significant reservations about whether current AI systems can actually carry their expected weight.
For ITDMs, this creates a specific planning tension. The instinct to accelerate AI adoption to stay competitive is sound. But if nearly half of AI leaders aren’t confident in agent autonomy, the skills that sit at the human-AI interface, specifically the ability to supervise, correct, redirect, and communicate with AI systems, become differentiating capabilities. The GoHumanize study frames this as “leadership and collaboration.” A more operationally precise framing for enterprise contexts would be: human oversight, escalation judgment, and cross-functional accountability.
What Developers and Technical Teams Should Take From This
For developers and technical practitioners, the GoHumanize study’s implicit message is uncomfortable but worth examining directly. If the skills most resistant to automation are social and managerial, where does that leave the software engineering profession?
The honest answer is: not where most developers expect. The work of writing boilerplate code, generating unit tests, and drafting documentation is already being absorbed by AI-assisted tooling. ECI Research’s 2025 AI Builder Summit data shows that two-thirds of enterprise AI leaders have already implemented multi-agent collaboration in live or pilot workflows. That’s not a future trend. That’s the current operating environment in leading enterprises.
What this means practically is that the developers who retain the most organizational leverage will be those who can define problems clearly enough for AI systems to act on them, review AI-generated outputs critically, and communicate tradeoffs to non-technical stakeholders. Those are fundamentally human skills. The GoHumanize framing of “coaching and mentoring” and “interpersonal skills” sounds soft, but the enterprise equivalent is the ability to translate between business intent and technical execution, a skill that has always been rare and is becoming more valuable, not less.
The Governance Gap Is the Real Risk
One dimension the GoHumanize study doesn’t address directly is what happens when enterprises reduce investment in the very human skills that AI systems most need to function safely. The study frames human skills as a protective buffer against automation. But there’s a second-order concern: if organizations interpret the study as validation to double down on human judgment while simultaneously accelerating AI agent deployment without adequate governance, they create a structural accountability gap.
ECI Research has observed that organizations with the most integrated teams, not the most advanced tools, achieve the highest FinOps maturity. The same principle applies to AI governance. Supervision, oversight, and accountability require the same interpersonal and organizational skills that the GoHumanize study identifies as hard to automate. The risk is not that AI replaces those skills. The risk is that leadership assumes AI has replaced the need for them.
Looking Ahead
The Skills Gap Will Widen Before It Narrows
Enterprise hiring practices and educational curricula have not yet adapted to the shift this study describes. Universities, as GoHumanize’s founder notes, still emphasize STEM and analytical training. That’s appropriate for many roles. But the demand signal for human judgment, communication, and oversight skills is strengthening as AI systems take on more execution work. Organizations that treat this as a distant HR problem rather than an immediate leadership and talent strategy issue will find themselves with capable AI systems and insufficient human infrastructure to govern them.
The transition to a workforce that includes both human and AI contributors is already underway in the enterprise. The question for the next 18–24 months is not whether it will happen, but whether organizations will invest in the human capabilities that make the human side of that workforce genuinely additive rather than merely residual.
The Measurement Problem
Enterprises should be cautious about translating a study like this directly into hiring frameworks or performance management systems. Skills like leadership, negotiation, and coaching are notoriously difficult to assess objectively, which is part of why they have historically been underweighted in technical organizations. As AI takes on more measurable tasks, the pressure to make human contribution more measurable will intensify. Organizations that invest now in structured frameworks for assessing and developing human judgment skills will be better positioned than those who wait for the market to define those standards for them.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
