Should an AI Agent Be Able To Fire You? Yes, And It's Likely To Go Well.

This week's Oracle layoffs sound like the future has arrived already.

Layoffs and redeployments are a fact of life, and now they’re becoming automated like never before.

Layoffs Used To Be A Big Thing

In the 1960s and 1970s layoffs were a big deal. Companies had long term relationships with their workers and it was unusual (or at least episodic) when a layoff took place.

Today, by contrast, this form of “redeployment” happens regularly, so it’s much more of an ongoing business process.

Historically Layoffs Were Personal

I lived through Lou Gerstner’s big layoffs at IBM (100,000 employees in 1990s) and the company took an honest, humane approach. Gerstner clearly explained IBM’s problems and sadly announced that for the first time in almost a century, IBM was letting people go. Folks got written letters, manila folders with transition packages, and managers had conversations.

I was sitting at IBM west coast HQ in San Francisco and watched almost an entire floor clean out their desks, weep and hug each other, and file out of the building. It was respectful, organized, yet traumatic for everyone.

Since then, this process has happened more and more.

In 1992 I went to work for fast-growing Sybase, a pioneer in client/server. We were a hot $100M company defining a new future but Microsoft and Oracle caught up. So after a massive acquisition (PowerSoft) the company stumbled, and in 1996 and again in 1998 we had big layoffs. Again there were lots of explanations but by the time the second one came we all felt stunned enough to expect it. (Eventually Sybase was sold to SAP.)

I went into online learning (startup company we sold) and became a VP at a .com company called DigitalThink. We went public, things where hot, and then suddenly we fell behind. And in 2001, right before 9/11, we had to downsize, and I was one of the ones to go. I was a VP at the time and the founder walked into the conference room and silently pushed a manila envelope my direction.

I asked “am I being laid off?'“ and he nodded his head. In a fit of anger I left the building and was so upset my wife had to come pick me up and drive me home.

All This Trauma Happens Regularly Now

As technology speeds up, markets move faster, and it’s now easier than ever for a business to “fall behind” or “make a bad bet.” So layoffs and restructuring is common.

Meta (Facebook) seems to do layoffs at least once or twice per year, often acquiring a few hot companies, placing big bets (the Metaverse, for example), and then later letting these people go. Despite the toll on employees the company has done very well, and I actually think this “dynamic talent reconfiguration” has served shareholders well.

And I believe this “rapid redeployment” is moving faster.

Here’s my perspective. Prior to 1980s or so, most employers (not all) felt a sense of responsibility to their workers. But sure enough, as time progressed, we all witnessed what I call a “decoupling” between employer and worker. Slowly, starting with Ronald Reagan’s firing of the air traffic controllers in the 1970s, there has been a steady trend toward this separation. Here’s my summary to listen to.

We can debate this 50 year “decoupling” trend later (it’s one of the highlights in my Superworker keynote at Irresistible 2026) but we’ve lost the idea that eliminating someone’s job may be trivial to the company but it’s a life-changing experience for the worker. And when it’s done poorly the company often suffers trauma and brand damage that lasts for years.

(The “Big Decoupling” over 50 years)

Blanket Layoffs Hurt Performance, Surgical Ones Don’t

I won’t cite all the research¹², but many studies have shown that companies who go through huge layoffs usually struggle. Not only does valuable knowledge walk out the door but the psychological trauma impacts the survivors. So maybe a better model is not to do it in one fell swoop, but rather think of it as a “dynamic rebalance.”

I’d simply mention the quote I heard from an aerospace HR leader long ago.

“Many of our engineers have been hired and laid off so many times they finally came to the conclusion that ‘I’m Never Working For That Company Again!”

So rather than sit around with HR and plan this massive restructuring every 3-5 years (look at HSBC, Oracle, Block, Atlassian), maybe this should be more dynamic?

Well with AI this all could change.

Why AI Agents Will Really Help

Let’s assume that any company, regardless of your success, goes through peaks and valleys - in product areas, markets, geographies, and customer segments. Nike is struggling to reinvent its brand; Starbucks and Chipotle are reinventing their consumer experience; and every tech company is trying to find it’s AI angle.

(Read Apple, the First 50 Years, to hear about how Apple almost went out of business a few times.)

When this takes place, why wouldn’t an AI Agent “see the trend” and come to management to advise and recommend a change?

The Agent (or Superagent, in our architecture) would see a low level of output per worker, it would know the job title and role skills in the job, and could recommend an internal redeployment, a job redesign, or simply a layoff?

I’m now convinced this is going to happen and it’s almost happening already. And we’ve been experimenting ourselves. Here’s an example.

Our AI Agent Galileo has been deeply trained on every major job title, key skills, and job family in business. We’ve built a training scenario where we tell Galileo about a simulated tech manufacturing company with R&D, manufacturing, sales, and distribution around the world.

We then tell Galileo about a major downturn in demand in one area of the business and we ask Galileo for advice. It takes the real job titles, roles, salaries, and org structures and we prompt:

“How would you reorganize, redeploy, or shrink the workforce to improve productivity and growth? Please consider internal redeployment, organizational flattening, reskilling, and layoffs in your plan.”

We give the agent a bunch of rules (average pay bands, employee tenure, implied employee skills, “time to productivity” for job changes and new hires, and desired span of control) as well as rubrics (culture, leadership values, company strategy) and believe it or not, it gives us options to redeploy people with amazing accuracy.

And there’s more. Because Galileo is trained on hundreds of real-world HR scenarios, it gives us change management options. It advises or discusses with us the idea of manager-led layoffs, redeployment by individuals or by leaders, and different modes of communication to workers. If you were using an agent like this in your company, you could fine tune it on all your internal behaviors, rules, and cultural values.

I’m not aware if any of our clients have used it for layoffs yet, but I know many have used it for reorgs, talent mobility projects and big reskilling initiatives. And since it’s connected to your company’s L&D programs and internal meetings and emails, it’s pretty smart about what people are good at and how well they’d fit in a new job.

This is Where “Layoffs” Are Going

While nobody wants to get “taken out” by a machine, compare this to the manual process today. I’ve sat through many meetings where people wave their arms and just “decide” that Joe, Sally, and Mike have to go but we’re keeping Bill, John, and Carol.

Or a leader says “we’re automating our testing process so lets get rid of the test group and keep two people” or “we have an automated scheduler now so let’s fire our admin and scheduling group” and on and on. Of course companies agonize through these decisions but there is no time to map each person’s skills to a new set of positions when you’re trying to downsize thousands of people.

A well trained AI Agent could do a personalized assessment, job match, and historic performance review (throw your 360’s and feedback forms in there and also have the AI read the person’s emails and meeting notes) and give every “displaced” person a highly personalized plan for transition.

Can we make this even more human? Yes of course - the “talent redeployment agent” can personalize each employee’s options, separation package, or options for reskilling.

AI, come to our rescue. Maybe you’ll make our lives better than we even thought.

How to Learn More

Our podcast (2-3 episodes a week now) is filled with stories, vendors, and tech discussions on this stuff.

Come to our amazing conference (Irresistible 2026) on June 8-10 at the beautiful USC campus to meet other leaders and discuss these issues.

Read our research and debate. We put ALL our research, case studies, and models in there - and we update it every day. It’s like the Bloomberg Terminal, Encyclopedia Brittanica, and HR Expert with new research, vendors, and case studies every day. . Get Galileo for yourself. Our entire 2026 AI Imperatives research is all in Galileo.

Dive into our Superworker studies.

Or call us or meet with us at the conferences around the world.

(Here is a clip from this week’s podcast digging into HR tech and AI announcements.)

1. Be Careful With The Layoffs https://joshbersin.com/2022/06/be-careful-with-the-layoffs/

2. How Layoffs Hurt Companies. https://knowledge.wharton.upenn.edu/article/how-layoffs-cost-companies/