Entry-Level Jobs Are Disappearing Fastest: Here Is What to Do Instead

Entry-level roles have been cut harder than any other tier in 2025. AI replaced the exact work that used to be how junior employees proved themselves: research, data entry, first drafts, basic analysis. The traditional career on-ramp is broken in finance, law, and content writing. What still works: AI trainer and evaluator roles, tech sales, customer success, trades, healthcare, and field work. For new grads, an ATS-optimized resume needs a strong skills section and detailed project descriptions since there is no work history to fall back on.

You applied to 40 jobs last month. You got two automated rejections and heard nothing else. Meanwhile, your university career center is still handing out the same advice it gave in 2019.

The entry-level job market is the most disrupted tier right now, and nobody is being straight about it.

The Numbers Are Real

Junior roles were not evenly affected across all sectors. The cuts were concentrated in specific areas, and the scale matters.

Finance took one of the hardest hits. Junior analyst and research associate positions at investment banks and asset management firms declined by roughly 35% from 2023 to 2025, according to tracked hiring data across Wall Street. Goldman Sachs, JPMorgan, and Morgan Stanley all cited AI automation of research compilation, data cleaning, and first-draft report writing in their workforce planning disclosures.

Content writing entry-level roles dropped by approximately 40%. The demand for junior blog writers, content coordinators, and marketing copywriters at agencies collapsed after widespread GPT-4 and Claude adoption in 2024. Agencies that once hired three junior writers now hire one senior strategist who directs AI output.

Junior legal research positions fell around 30%. Law firms piloting Westlaw AI, Harvey, and Casetext found that first-year associates spent 60-70% of their time on work that these tools could now do in minutes.

The pattern is clear. These were not random cuts.

Why This Particular Tier Got Hit

Entry-level work was built around a specific premise: you do the rote, repetitive, learnable tasks while senior people check your output and you absorb how the business works. Over time, you build judgment by doing volume.

AI is very good at rote, repetitive, learnable tasks.

The learning-by-doing path depended on those tasks existing. When AI handles the bulk research, the data cleaning, the first draft, the document summarization, junior employees lose the work that was supposed to teach them the craft. It is a genuine career ladder problem, not just a job availability problem.

Senior roles are largely intact. Mid-level roles have become more productive with AI tools but remain staffed. The gap opened specifically at entry level, where the training-wheels work used to live.

What Entry-Level Still Looks Like in 2026

The good news is that new openings exist. They just do not look like the old entry-level jobs.

AI trainer and evaluator roles. Scale AI, Turing, Surge AI, and dozens of internal AI teams at large companies are hiring people to review model outputs, flag errors, label edge cases, and write evaluation rubrics. These roles scale with AI deployment, not against it. Entry-level pay runs $45,000 to $65,000 annually, with remote options common. The work requires judgment, attention to detail, and domain knowledge in whatever area the model is being trained for.

Tech sales and customer success. Enterprise software companies are hiring aggressively at the entry level because selling and retaining SaaS customers requires human relationship management. An AI cannot build rapport with a procurement team or talk a client down from canceling. Account development representative roles at companies like Salesforce, HubSpot, Outreach, and hundreds of smaller SaaS firms have remained consistently open. Base salary plus commission for entry-level reps typically lands at $55,000 to $75,000 total compensation in year one.

Healthcare and allied health. Nursing, physical therapy, occupational therapy, medical assisting. These roles require physical presence, real-time patient assessment, and licensed judgment. They were not disrupted by AI in 2025 and will not be disrupted soon. Entry-level clinical roles are in shortage in most US metro areas and much of Europe.

Skilled trades. Electricians, HVAC technicians, welders, plumbers. Decades of pushing everyone toward four-year degrees created a genuine shortage. A first-year electrician apprentice in the US earns more than many entry-level office jobs, and the work cannot be offshored or automated in any near-term sense.

Field operations and logistics. Warehouse roles requiring physical problem-solving, route delivery, quality inspection, and facilities management are staffed at entry level. These are not always the roles new graduates have in mind, but starting here with a plan to move into operations management or supply chain strategy is a real career path that exists right now.

The AI-Adjacent Entry Strategy

One approach that is working for 2026 graduates: position yourself as someone who makes AI-augmented teams run better, rather than competing for the tasks AI replaced.

What this looks like in practice is identifying where human oversight of AI output is required and presenting yourself as the person who provides it. Legal teams using Harvey still need someone who reads the AI-generated research and catches the errors before it goes to a partner. Content operations using AI writing tools still need someone who edits for brand voice and catches factual mistakes. Data teams using AI analysis tools still need someone who understands whether the output makes sense.

The framing in a cover letter is direct: “I understand how these tools work, I know where they fail, and I can provide the review layer your team needs.”

This positions you for a role that exists in 2026, rather than a role that existed in 2022.

Building a Portfolio When You Have No Work History

The biggest practical problem for new grads right now is that they have no professional work history to point to, and the entry-level jobs that used to build that history are fewer. The answer is to build it yourself before you apply.

For technical fields, GitHub is the primary portfolio surface. Three or four well-documented projects that solve real problems are more persuasive than a GPA at most tech companies. The project does not need to be original. A clean implementation of a known algorithm with good documentation, a data analysis of publicly available data with clear methodology, or a small web app that does something specific all count as evidence of competence.

For business and operations roles, case studies work. Take a publicly available business problem, work through it in a structured document, and publish it. McKinsey makes its case interview prep materials public. Consulting firms publish anonymized case studies regularly. Working through three of them and posting your analysis on a personal site or LinkedIn demonstrates reasoning skills to hiring managers who cannot rely on intern experience or part-time work to evaluate you.

For legal and compliance careers, volunteer work with organizations that need legal research support is real experience. Law school clinics, nonprofit advocacy organizations, and legal aid societies are often willing to take on organized, careful research help from people in paralegal or pre-law roles.

Freelance work at any scale counts. Even two or three paid projects through Upwork or direct client work gives you a “clients” section on a resume rather than a blank space.

The ATS Problem for New Grads

Here is something career centers rarely explain clearly: applicant tracking systems score resumes in part based on years of experience. A new grad competing against someone with two years of relevant work will typically score lower on that dimension. You cannot fix that. What you can fix is every other dimension the ATS scores on.

For new grads, the resume sections that carry the most weight with ATS systems are the skills section, education, certifications, and project descriptions. These need to be written with the same keyword discipline that experienced candidates apply to their work history section.

Practical steps:

First, read the job description carefully and identify the specific skills and tools it mentions. Then verify that your resume uses the same language. If the job says “data visualization” and your resume says “charting,” the ATS may score them separately even though they mean the same thing. Match the exact phrasing when you have the competency.

Second, your project descriptions should function like work experience bullet points. “Built a Python data pipeline that processed 500,000 rows of public census data to identify housing cost trends by county” is a better project description than “Python data analysis project.” The ATS reads the specificity and the keywords. So does the hiring manager.

Third, certifications matter more at entry level than they do later in your career. Google Data Analytics, AWS Cloud Practitioner, Salesforce Admin, CompTIA Security+, the CFA Level 1. These are not substitutes for experience, but they are signals that score well with ATS systems when the alternative is a blank work history. Many are available for free through Google Career Certificates, Coursera financial aid, or employer-sponsored programs.

Check your resume score against actual job descriptions you are applying to using an ATS resume checker. The feedback on which skills are missing or which keywords are not matching is more actionable than general advice about formatting.

Avoiding the Resume Black Hole

Most entry-level applicants are applying to roles through job boards, where their resume competes against hundreds of others and the ATS filters happen before any human sees the document. The application-to-response ratio at popular job postings on LinkedIn and Indeed for entry-level roles is often below 2%.

Two approaches improve those odds.

First, target companies with fewer than 100 employees. Small companies post on job boards less frequently but hire from direct outreach more often. A short, specific email to a hiring manager at a 30-person company explaining what you can contribute gets read more often than the same email sent to a Fortune 500 HR inbox. The competition is smaller and the decision-making is more personal.

Second, reach behind the job posting when possible. Find the person who would be your direct manager on LinkedIn. Send a short note that is not a form letter. Reference something specific about their work or the company’s recent projects. This does not work every time. It works often enough that it is worth doing for every role you are genuinely interested in.

For details on why the standard online application process fails at such high rates, see why your resume gets no response in 2026.

Your Next Steps

The entry-level market is harder than it was three years ago. That is a real fact, and pretending otherwise is not helpful. But harder does not mean closed.

The graduates finding roles in 2026 are doing three things differently: they are targeting the categories where entry-level hiring still exists, they are building portfolios instead of waiting for experience to arrive, and they are writing resumes that are optimized for ATS systems rather than formatted for human readers who may never see them.

Check your resume’s ATS score against the roles you’re targeting. For a new grad, the difference between a resume that passes ATS filtering and one that does not is often a skills section and a few well-chosen keywords. That gap is fixable.

For context on how AI is changing hiring more broadly, see AI resume screening in 2026.