AI Layoffs and the Severance Crisis

Something shifted and nobody sent a memo

If you work a mid-level white-collar job — operations, HR, marketing, finance support, project management, content, customer success — you've probably noticed something in the last year or two. Teams are getting smaller. Headcounts aren't being replaced. New hires are coming in junior and cheap, or not coming in at all. The org chart got flatter and nobody announced it.

Not a recession. Not a hiring freeze that bounces back in Q3. This is structural — the way companies staff themselves is changing permanently, and AI is doing most of the pushing.

Not the dramatic, Hollywood version of AI where a robot sits in your chair. The quiet version. The version where your company subscribes to a platform that handles 60% of what three people on your team used to do. The version where your manager's manager looks at a dashboard and realizes the department can run with four people instead of seven. The version where the layoff email says "restructuring" and never once mentions the word automation.

The numbers nobody's organizing

What makes this moment different from previous layoff waves? Scale and invisibility. When manufacturing jobs moved overseas in the 90s and 2000s, you could see the factory close. When retail contracted, you could see the stores go dark. White-collar AI displacement is almost invisible because it doesn't happen in one dramatic cut — it happens in rolling reductions, quiet non-replacements, and strategic restructurings that are carefully kept under the WARN Act threshold.

The data exists in fragments. An earnings call where the CFO mentions "efficiency gains from AI implementation." A LinkedIn post from someone who got walked out on a Tuesday. A Reddit thread where 40 people share what happened at their company this quarter. A WARN filing in Ohio that lists 87 positions eliminated in "operations support." Put it together and a pattern emerges. Pull any one piece out and it's just an anecdote.

Nobody is aggregating this into something usable for the actual people affected. Journalists cover the big ones — Google cuts 12,000, Meta cuts 11,000 — but the mid-size company that quietly let go of its entire marketing ops team doesn't make the news. Those people still need to negotiate their severance and figure out their next move.

Severance practices are stuck in 2015

OK so the part that actually matters if you're reading this because you got the call — or you think it's coming.

Severance packages haven't evolved. At all. The standard formula most companies use — one to two weeks of pay per year of service — has been the default for decades. It was designed for a labor market where a mid-career professional could reasonably expect to find comparable work within a few months. That assumption is breaking down.

If your role was eliminated because a platform now does 70% of what you did, finding "comparable work" means competing for a shrinking pool of similar positions — along with everyone else whose role just got automated. The job search takes longer. The salary often comes in lower. The severance package that would have been adequate five years ago doesn't cover the gap anymore.

And companies know this. They know the job market for displaced mid-level workers is tightening. But severance practices haven't adjusted because there's no external pressure forcing them to. There's no federal severance standard. There's no regulatory requirement to account for the reason behind the layoff. A company can replace your entire department with AI tools, hand you four weeks of pay, and call it generous.

The infrastructure that doesn't exist

When you get laid off, here's what the internet offers you:

Job boards. LinkedIn, Indeed, Monster. These assume you know what you want and can find it. They're fine if your skills transfer cleanly. They're useless if your entire function is contracting.

Law firm blogs. Every employment attorney in America has a page called "Is My Severance Package Fair?" They all say the same thing: it depends, call us. The page exists to generate leads, not to give you a benchmark.

HR publisher articles. Written for employers, not employees. They explain what severance is. You already know what severance is — you're holding the offer letter.

Reddit threads. Ironically, the most useful resource. Real people sharing real numbers. "Got laid off after 8 years at a mid-size SaaS company, offered 4 weeks, negotiated to 12 weeks plus 3 months COBRA." The problem is it's scattered across hundreds of threads, completely unstructured, and impossible to search in any meaningful way.

And... that's basically the whole landscape. Job boards for finding work, law firms for legal questions, Reddit for actual peer data. Nothing in between. Nothing that organizes what people actually received, compares it against your situation, and tells you plainly whether you're getting a fair deal or getting lowballed.

Which is why SeveranceScore exists.

What we're building and why

SeveranceScore is a benchmarking tool. You enter your role, tenure, industry, company size, and what you were offered. It tells you where your package falls compared to what people in similar situations report receiving. Below typical, typical, or above typical — with the actual numbers behind the comparison.

The tool is backed by three layers of data. Published industry benchmarks — the standard ranges that HR consultants and employment attorneys reference. Structured data extracted from public forums where workers voluntarily shared their severance terms. And anonymous submissions from people who use the tool. Every submission makes the dataset better and the comparisons more accurate.

There's no account to create. No email to enter. No data that traces back to you. You get your result, you get a few practical next-step suggestions — not legal advice, just things worth knowing — and you leave better informed than you arrived.

The content around the tool exists because people don't just search for a score. They search for "is 4 weeks fair after 10 years" and "should I negotiate my severance" and "can I collect unemployment while getting severance in New York." Those questions deserve straight answers backed by real data, not 2,000 words of filler that ends with "contact our office for a free consultation."

The AI context changes everything

Timing matters here. A lot. Previous layoff waves — dot-com bust, 2008, early COVID — were cyclical. Companies cut, the economy recovered, the jobs came back in some form. Workers who toughed it out or retrained could re-enter roughly the same labor market.

AI-driven displacement is structural. The roles aren't coming back. They're being permanently replaced by systems that get cheaper and more capable every quarter. That means the old playbook — take the severance, update the resume, apply to similar positions — has a shorter shelf life than people realize.

I'm not trying to scare anyone. People will adapt — retrain, pivot, build new skills, move into adjacent roles. But they'll do all of that better from a position of financial stability than from panic. And financial stability after a layoff starts with one thing: knowing whether your severance package is actually fair and having the leverage to improve it if it isn't.

A few extra weeks of severance isn't just money. It's time. Time to think clearly about what's next instead of panic-applying to whatever's posted. Time to explore whether consulting or fractional work makes sense. Time to retrain without burning through savings. The difference between four weeks and twelve weeks can be the difference between a thoughtful transition and a desperate one.

The age dimension nobody talks about

There's another layer to this that gets whispered about but rarely said directly. The mid-level roles most vulnerable to AI automation are disproportionately held by workers in their 40s and 50s. People with fifteen or twenty years of experience in a function that's being compressed. People whose salaries reflect two decades of tenure and raises — salaries that make them expensive relative to what an AI system costs.

These workers have legal protections. The Age Discrimination in Employment Act covers everyone 40 and older. The Older Workers Benefit Protection Act gives specific rights around severance agreements — 21 days to review, 7 days to revoke, 45 days if you're part of a group layoff. But protections only work if you know they exist and if you exercise them. Most people don't, because nobody tells them in plain language what those rights actually mean in practice.

It's also the demographic least likely to negotiate. Employment attorneys consistently report that older workers accept initial severance offers more often than younger ones, frequently leaving significant money on the table. There's a generational tendency to feel grateful rather than strategic. That instinct is understandable. It's also expensive.

If you're over 40 and reading this because you just got laid off or you see it coming — take the full 21 days. Don't sign anything under pressure. And check what people in your situation actually receive before you decide what's fair. That's literally why we built the scoring tool.

What happens next

The layoffs will accelerate. Nobody wants that to be true but the economics are pretty irresistible — companies have access to tools that replace $80,000-a-year roles for a fraction of the cost. Already happening, just slowly enough that it doesn't make headlines every day.

As it accelerates, the need for worker-side infrastructure gets more urgent. Not more job boards. Not more generic career advice. Tools that give people leverage and information at the exact moment they need it — standing in the HR office, holding a severance agreement, trying to figure out if they should sign.

SeveranceScore is one piece of that infrastructure. A small one, honestly. But it's the piece that doesn't exist yet and should. If you're reading this and you have a severance offer in front of you, go score it. If you're reading this and you think the call might be coming, bookmark this and come back when it does. Either way, you deserve to know what fair looks like before someone asks you to sign away your rights.