I've been paying close attention to the labor market lately, and something keeps nagging at me.

AI tools are quietly raising the output ceiling at work without raising pay to match. Hiring teams will tell you their processes are running smoothly, but candidates are navigating something that feels nothing like that. It now takes longer to fill a skilled trades role than a software job.

There’s still a chance for you to win your $50 Amazon Gift Card. 3 questions, 2 minutes, and 1 chance to win a $50 Amazon gift card. Your answers will shape exactly what’s created next. Building Cocoon Careers Survey

Events for You:

I've been spending a lot of time recently learning the inner workings of Claude, building out my own workflows, automating tasks I used to do manually, and generally trying to understand what this technology actually does when you push it. In the middle of all that, something clicked. Not about AI, but about me and the expectations forming around me.

If AI saves you two hours a day, your employer doesn't necessarily give you those two hours back. New research from Culture Amp makes this uncomfortably visible. While 96% of C-suite leaders expected AI to boost output, 77% of employees said AI tools actually increased their workload. The logic, rarely stated out loud, goes something like “If your tools got faster, your output ceiling just went up”. This is the AI tax. It doesn't come out of your paycheck. It comes out of your time, your energy, and eventually your engagement.

Agentic workflows are no longer experimental. Companies are running multi-step automations, deploying AI agents across functions, and redesigning how work moves through a team. That's genuinely useful. It's also creating a new invisible expectation that, because more can be done, more should be done, by you, now, at the same salary.

U.S. worker engagement has dropped from 36% in 2020 to 31% in 2024, with an estimated 8 million fewer workers actively engaged over five years. That's not a coincidence landing in the same era as mass AI adoption.

So what do you do with that?

Efficiency was never supposed to mean infinite availability. The professionals who understand that now will be the ones who still have something left to give in three years.

Was this newsletter forwarded to you? Subscribe here to get weekly insights that help you navigate your career with more clarity and confidence.

There's a number in Starred's 2026 Hiring Benchmarks Report that should make every manager pause. Hiring managers rated their satisfaction with the recruitment process at +73. Candidates rated their experience at +17. That 56-point gap shows blind spots and it's been growing.

The report, built on over 2.5 million hiring experiences, finds that talent acquisition teams worked harder than ever in 2025. Processes became more structured, internal alignment improved, and decision-making felt more data-driven. The inside of hiring looked increasingly mature, but the outside looked like a black box.

What's easy to miss is where candidates actually form their opinion. Nearly 90% of surveyed candidates were people who were rejected. Of those, close to 70% were filtered out at the application stage, often before any human ever touched their file. That rejection email, or that automated silence, is where most people decide what they think about your company. 

When a candidate gets screened out by an automated system with no explanation of what was assessed or why, they don't conclude that the process was efficient. They fill in the gap themselves. Across nearly 70,000 open-text comments, the top driver of dissatisfaction wasn't rejection. It was the absence of clarity around it.

For mid-career professionals, this reframes something practical. The companies treating early-stage communication as an afterthought are leaving their employer brand in the hands of whoever wrote the auto-reply. For those of you managing hiring processes or advocating for your teams inside organizations, the report's finding is a useful pressure point. Internal alignment is only half the work. What candidates experience is the other half, and most teams aren't measuring it.

The teams that consistently outperformed the benchmark weren't using more technology. They were using it differently by keeping humans in the decision seat while letting automation handle the admin, and treating communication at every stage as a metric worth measuring. The gap between +73 and +17 doesn't close by accident.

Somebody buried the real story on AI.  While everyone spent the last two years debating which desk jobs a chatbot would replace next, something quietly flipped in the labor market. It now takes longer to hire an electrician than a software developer. Not slightly longer. We're talking 56 days to fill a skilled trades role versus 54 days for a desk-based professional, according to a new Randstad USA analysis of more than 150 million job postings between 2022 and 2026.

Two days doesn't sound like much until you understand what's behind it. Skilled trades demand is growing three times faster than professional roles right now.

Every data center being built, every power grid being upgraded, every cooling system going into a server farm requires skilled hands to do it. As Randstad's chief economist put it, AI can't build its own infrastructure. So who does? Electricians. Welders. Industrial technicians. The people we told to consider a different career path.

That's the other side of this story. For every 100 young workers entering the trades today, 102 are leaving. Decades of steering young people toward four-year degrees had consequences, and we're measuring them now in days-to-hire. We've touched on pieces of this before, the rise of new-collar jobs, apprenticeships making a comeback, employers investing in workforce pipelines, but what the Randstad data does is put a number on something that was always true and rarely said out loud. The most AI-adjacent work in the country doesn't happen at a keyboard, It happens on a job site.

For you, that's a positioning question worth sitting with. If your work touches operations, facilities, infrastructure, or anything at the intersection of physical systems and technology, your market value looks different today than it did two years ago. The window to be early on this read is still open. But it won't be for long.