Clara Shih spent two decades leading AI products at Salesforce and Meta. She watched agentic AI compress product development cycles and shrink team sizes-then started taking down her own entry-level job postings. At the same time, young family members with strong resumes couldn't get hired. That collision led her to leave Meta this spring and launch the New Work Foundation, a nonprofit aimed at closing what she calls the "information asymmetry" between people who work in AI and everyone else.
The foundation runs three tools for young job seekers: a media arm called "Dear CC" that publishes advice from managers and Gen Z workers; Field Report, a data tool showing hiring trends, salary data, and exposure rates for different majors and occupations; and Game Plan, a platform that helps job seekers build a search plan and connects them with mentors and peers.
What drove a Big Tech leader to act
Shih said agentic reasoning models changed her calculus. Large language models moved from answering questions to handling multi-step cognitive tasks. "When it really hit home for me was just talking to the people who were using our business agent products [at Meta], and deploying agents internally to accomplish everything from marketing to privacy policy review, our sales strategy, product roadmap, prototyping-pretty much every aspect of building, delivering, and taking a product to market."
She described a "perfect storm" of three signals: customers deploying business agents and needing fewer people, her own team using AI agents internally, and family members who had done everything right-internships, campus leadership-yet couldn't land jobs. Their messages to her were direct: job offers rescinded, interviews conducted entirely by AI, no human recruiter on the line.
The data backed up what she saw. Entry-level hiring has fallen, and Shih can point to specific roles she eliminated or changed the experience requirements for. "That was because of AI," she said.
The employer side of the problem
Shih is blunt about why companies are cutting entry-level roles first. Businesses have spent heavily on AI pilots, most of which haven't shown ROI. They don't know how to upskill their workforce. "It's often easier just to cut entry-level jobs and not hire people who haven't started yet than to do layoffs, so that's where they start."
The New Work Foundation wants employers to think longer-term about talent pipelines. The nonprofit is in early talks with a few employers willing to make large entry-level hiring commitments, with the foundation vetting candidates who have the right AI skillset and mindset. In February, IBM CHRO Nickle LaMoreaux announced the company was tripling entry-level hiring-a move that drew attention.
For product teams watching AI reshape how work gets done, the shift Shih describes is already visible in AI for Product Development workflows. The question is whether companies treat entry-level roles as expendable or as the start of a pipeline that needs rebuilding.
Why this matters for product development professionals
Shih's experience cuts two ways for product teams. First, AI agents are now handling tasks that used to belong to junior hires-market research, prototyping, sales strategy, roadmap drafting. That changes what "entry-level" means. Second, the same pressure is reshaping how product managers build their own teams. If you're a PM or product leader, the entry-level roles you post today may not exist in a year, and the skills you screen for need to shift accordingly. The AI Learning Path for Product Managers covers the practical side of that transition: what to automate, what to keep human, and how to staff for both.
Shih's closing point is simple. "When you have knowledge about how something could play out and you have the skills and experience that are required to do something about it, then you have to do something about it." For product leaders seeing the same signals, the decision is whether to wait for the data or start adapting now.
Your membership also unlocks:
