The 20 New Agentic AI Jobs Box McKinsey And LinkedIn All See Coming (Photo by Smith Collection/Gado/Getty Images).
Getty Images
The dominant headline about Agentic AI and work is the one about jobs disappearing. The data tells a second story that gets far less attention.
LinkedIn's January 2026 analysis put a number on it. AI has already added more than 1.3 million new roles, titles like AI Engineer, Forward-Deployed Engineer, and Data Annotator, alongside over 600,000 AI-enabled data center jobs. LinkedIn calls this the new-collar era, a workforce that blends knowledge work, advanced technical skills, and distinctly human strengths. AI Engineer now ranks as the single fastest-growing job title in the United States for the second year running.
Box gave the trend a face this week. Responding to a New York Times piece on its AI-driven hiring, CEO Aaron Levie described positions that did not exist at the company two years ago, spread across engineering, IT, and go-to-market: roles that evaluate new AI models, internal automation engineers who rewire how the company works, and professional services teams running a forward deployed engineering motion to get customers live faster.
His point was that Box is not the exception. It is the early indicator, and the same shift is happening inside its customers.
The institutional research agrees.
The World Economic Forum's Future of Jobs Report projects 170 million new roles created by 2030 against 92 million displaced, a net gain of 78 million jobs, with AI and machine learning specialists among the three fastest-growing occupations worldwide. McKinsey frames the destination as the agentic organization, where generalists become orchestrators of AI agents and entirely new roles emerge to supervise, coach, and govern them.
So the question is not whether AI creates jobs. It is which jobs, and which of them signal a company that has actually restructured rather than simply announced a strategy.
Here are the twenty that matter most.
The 20 Future Jobs in Agentic Ai both in Technical Areas and Beyond
Sandy Carter
Embeds inside a customer to build and tune agentic workflows for their stack. Palantir originated it; OpenAI, Anthropic, Google, and EY are all hiring for it, and LinkedIn names it among the three roles driving the 1.3 million figure.
The fastest-growing title in the country, with 75,000 of LinkedIn's 639,000 new AI postings. Builds and runs AI products, agents, and large language models. Often relabeled Applied AI, GenAI, or LLM Engineer.
Box's eval new AI models hire. Builds the test suites that catch hallucinations, regressions, and bias before they reach a customer. A permanent discipline now that models drift with every update.
Designs the systems that feed a model the right information at the right moment, from retrieval pipelines to memory. As agents take on multi-step work, this role decides whether they succeed.
Designs autonomous multi-agent systems where models orchestrate tools and chain complex actions. The engineering core of agentic commerce, support, and operations.
Owns enterprise-wide AI strategy and ethical implementation. An IBM survey found one in four companies now have one, with two-thirds expecting most companies to follow within two years. Smaller firms hire fractional CAIOs.
A new role that is growing is around AI Ethics and governance.
getty
Keeps AI systems transparent, accountable, and aligned with the EU AI Act and state laws, managing bias, fairness, and privacy risk. Already mandatory in regulated industries.
Maps to Box's internal automation hires. Runs adoption programs, trains employees, and tracks whether tools are actually being used. McKinsey calls this the agent coach.
Stress-tests AI systems adversarially, probing for jailbreaks and unsafe behaviors before deployment. As agents gain the power to take actions, the cost of skipping this rises.
McKinsey's T-shaped expert who reimagines workflows and safeguards quality. Designs where humans and AI split the work, and what happens when model confidence drops. A natural pivot for consultants and analysts.
Here is where the bigger employment story lives. McKinsey is blunt: the largest new category is not engineers, it is AI-augmented frontline workers in sales, service, HR, and operations.
These ten roles will outnumber the technical ones, and most require no code.
LinkedIn's number two fastest-growing role, behind only AI engineer and far less technical. Translates capability into business cases and sequences the rollout. A direct path for consultants and business leaders.
Owns the roadmap for AI features and agentic products, sitting between what the model can do and what the market will pay for. As every product becomes an AI product, one of the most sought-after non-engineering roles.
McKinsey's M-shaped supervisor who orchestrates agents and the hybrid workforce. KPMG argues agents belong on the org chart with reporting lines, managed like human talent. The person running that digital team is a manager, not an engineer.
Teachers for AI, labeling the data models learn from and tuning outputs through human feedback. LinkedIn names data annotator among the three roles behind the 1.3 million. One of the most accessible on-ramps into AI work.
The voice of the agent. Shapes how AI talks to customers and employees, the tone, the flow, the fallback when it is unsure. A writing and experience discipline, not a coding one.
A new HR function. KPMG argues HR should manage agent performance and engagement like human counterparts. Owns workforce planning that counts both people and agents, and rebuilds career paths for a hybrid team.
Owns IP, data rights, and compliance with the EU AI Act and a growing patchwork of state laws. When an agent takes an action that costs money, this is the person who decided in advance who is responsible.
McKinsey's largest category: frontline sellers who spend less time in systems and more with humans while agents handle research and follow-up. The role changes shape rather than disappearing.
Captures the demand Box described at its customers. Drives adoption and proves outcomes inside accounts, not just the sale. Customer success rebuilt around agent deployment and expansion.
Independent verification that AI systems meet standards like ISO 42001, the first international standard for AI management systems. A natural pivot for risk, audit, and compliance professionals.
Each of these twenty roles is evidence of a company that changed its org chart, not just its press release. They get hired only when leadership commits to running the business through agents, which is why KPMG found 70% of large companies planning to restructure within three years.
Ten of these are not engineering jobs at all. They are strategy, product, HR, legal, sales, and assurance roles that did not exist in their current form two years ago. The hiring data shows it is becoming an operating model. The jobs AI eliminates will keep making headlines. The Agentic AI jobs it is creating will quietly decide which companies still exist in five years.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。