AI Is Delivering Real Returns for Companies & Eliminating Real Jobs at the Same Time
A new report from Snowflake and Omdia surveyed more than 3,400 companies worldwide about what generative AI is actually doing inside their organizations right now, not what they plan to do, not what they’re piloting, but what has already happened. The findings are some of the most concrete data we’ve seen on AI’s real-world workforce impact.
The headline number: 92% of early AI adopters report a positive return on their investment. Among those who’ve measured it, the average return is 49%, meaning $1.49 back for every dollar spent, up from 41% just a year ago. Companies are making money on AI; that part of the debate is largely settled.
The workforce impact is where it gets more complicated and more important for anyone paying attention to the job market.
The teams experiencing the most job losses due to AI over the past year are operations (40%), customer service (37%), and data analytics (37%). By seniority, 63% of respondents reported job losses at the entry level, and middle management was next at 46%. That entry-level figure aligns with Stanford research cited in the report, which found that early-career workers aged 22 to 25 in AI-exposed roles experienced a 13% relative decline in employment over the same period.
For anyone arguing that AI-driven job displacement is theoretical, this report makes that position very hard to hold. These are companies, thousands of them, describing what has already taken place inside their organizations.
The picture isn’t entirely bleak, though, and it’s worth saying that clearly. When companies were asked whether AI created jobs, cut jobs, or both, 42% said it only created jobs. Eleven percent said only jobs were lost. Thirty-five percent said both happened, and among that group, 69% said the net impact was positive. The biggest net winners for job creation were cybersecurity, up 21 points net positive, IT operations at 16 points, and software development at 12 points. The biggest net loser was customer service, where 37% reported cuts and only 15% reported gains, a 22-point gap that reflects how directly AI can replace that function.
One more number worth sitting with: 69% of respondents said it’s likely that AI agents will take over parts of their own job within 12 months. The self-awareness there is notable. The question for workers and employers alike isn’t whether this is coming. It’s whether they’re positioned for it.
Less Than Half of Workers Think Their Company’s Outlook Is Positive
Glassdoor released its February Employee Confidence Index this morning, and the number moved in the wrong direction again. The share of employees reporting a positive six-month business outlook dropped to 44.3% in February, down from 45.9% in January.
Less than half of American workers currently believe their company is heading in a good direction. That’s a retention warning as much as it is a sentiment reading. Disengaged employees don’t necessarily leave right away, but they underperform, they disengage, and eventually they do leave. If your business is healthy, your employees should know it. If it isn’t, but you have a plan, they should know that too.
The sector breakdown is telling. Information technology saw the biggest year-over-year confidence drop of any industry, down 7.1 points, which is striking given the scale of AI and data center investment happening in that space. The lowest overall confidence readings belong to restaurants and food service, manufacturing, and government workers.
The seniority data is where employers should pay the most attention. Senior-level employee confidence has fallen more than four points over the past year, the steepest decline of any group, and that trend has accelerated since economic policy uncertainty spiked around May 2025. Senior employees have more visibility into budget decisions, restructuring conversations, and strategic direction than anyone else in an organization. When they’re the most worried, that’s not a coincidence. It’s a signal.
For organizations considering how to support and retain employees amid uncertainty, confidence below 50% is the kind of number that demands a response, not a wait-and-see approach.
ADP’s Weekly Jobs Pulse Shows Steady but Modest Private Hiring
ADP’s weekly jobs pulse released this morning shows private employers added an average of 15,500 jobs per week for the four weeks ending February 21, holding steady after five consecutive weeks of strengthening. For context, the four-week average back in early December was just 2,500 jobs, so the trajectory has improved meaningfully.
That said, 15,500 jobs per week translates to roughly 62,000 per month, which is consistent with ADP’s February monthly report and not a number that signals strong labor market momentum. As Pete noted on today’s episode, those gains remain heavily concentrated in healthcare and construction, two sectors carrying a market that most other industries are sitting out.
The weekly pulse is most useful as a directional indicator rather than a definitive jobs count. The direction is modestly positive, but the foundation is narrow. Talk to our team if your organization is trying to hire in one of the sectors that isn’t participating in that growth right now, because the staffing strategies that work in a broad hiring market don’t always translate to a concentrated one.
Find the perfect fit for your team.
Speak to one of our recruiting experts today.
Frequently Asked Questions
According to a Snowflake and Omdia survey of more than 3,400 companies, customer service has seen the widest gap between job losses and gains, with 37% of companies reporting cuts and only 15% reporting gains. Operations and data analytics roles also saw significant losses. Entry-level workers have been hit hardest by seniority, with 63% of companies reporting job losses at that level.
Yes. The same survey found that 42% of companies said AI only created jobs, with no losses reported. Cybersecurity saw the strongest net job creation, 21 points more likely to report gains than losses. IT operations and software development also showed strong positive net employment impacts. Among companies that saw both job creation and job loss, 69% reported the net impact was positive.
Among companies that have measured it, the average ROI on generative AI investments is 49%, meaning $1.49 returned for every dollar spent. That figure is up from 41% a year ago, and 92% of early adopters report a positive return overall.
The share of employees reporting a positive six-month business outlook fell to 44.3% in February, down from 45.9% in January. Less than half of American workers currently believe their company’s outlook is positive. Senior-level employees showed the steepest year-over-year confidence decline of any group, dropping more than four points.
View Full Transcript
Snowflake/Omdia: What 3,400 companies say AI has already done to their workforce
Welcome back to Cornering the Job Market. Today’s headlines include Glassdoor’s Employee Confidence Index that came out this morning and ADP’s weekly jobs pulse. But first, Snowflake just published a survey of more than 3,400 companies worldwide focused on what generative AI is actually doing inside their organizations. And this isn’t about what they are planning to do or might do, it’s what they’ve already done. And the headline number, well, it jumps out at you. 92% of early adopters say they’re seeing a positive return on their AI investments. Among those who’ve measured it, the average return is 49%. That’s$1.49 back for every dollar spent, which is up from 41% a year ago.
So companies are making money on AI. You see a lot of reports saying they aren’t. This one very clearly says that they are. So it’s not just what you feel is happening or think might be happening. These are companies, lots of them, actually telling us what is going on inside their organizations. Also, middle management, that was hit the next at 46%. We’ve seen a lot of talk of that too. I know that Alex Carp recently said that middle management’s going to probably get hit the hardest as a result of AI. Seems that that’s happening already. So for the crowd that thinks AI is just much ado about nothing, there really isn’t going to be job loss, that’s becoming harder to defend. We continue to see reports on practically a weekly basis now telling us that it’s not theory, it’s practice. This is what’s taking place.
And that 63% tracks with Stanford research cited in this report. It shows that early career workers ages 22 to 25 in AI exposed jobs saw a 13% relative decline in employment. This report isn’t all bad news for workers because when companies were asked whether AI created jobs, cut jobs, or both, 42% said only jobs were created. So that is a surprise to me to see, a very positive one. 11% said only jobs were lost, and 35% said both happened. And among those who saw both, 69% said the net impact was positive. So this is really good to see. It’s rare data that’s good to see about job loss as it relates to AI. So I think more than one thing can be true.
I tend to think and say that a lot in this context where, yes, AI is impacting jobs, but it’s not impacting all jobs, and it is creating opportunities for those in certain areas. And that’s where I believe we’re going to uh how we’re going to see this play out over time. There’s going to be winners and losers in this, but for everyone individually, it is incumbent upon you to take advantage of what AI can do to elevate your work, whatever that might be. That’s really important here. And also, as it relates to career development, something that I like seeing is I have a son who’s a cybersecurity master’s student right now. The biggest winner for net job creation was cybersecurity. 21 points more likely to report gains than losses. I love seeing that. Also, IT operations was 16 points net positive, and software development was 12 points net positive.
So we see that AI is being used heavily in these areas, but it’s an example of where opportunities are being created. So interesting dynamic going on there. The biggest net loser was customer service, 37% reported cuts and only 15% reported gains. That’s a 22-point gap, which is significant. That’s the widest of any department they measured. And I think the difference is with customer service, AI can directly replace what’s done. That is very clear. That’s happening, I won’t say at scale yet, but it’s happening on an increasing basis. If your job involves answering questions that a chatbot can handle, this is happening. That shows really good self-awareness of what’s happening. And once again, the what’s important here is to not be surprised by it. Be very aware of how your company is adopting AI. Be aware of how you can use it to improve your work, whatever that might be.
Glassdoor: When less than half your workers are confident, that’s a retention problem
And the next story Glassdoor just released its employee confidence index for February. It comes out every month around this time. And unfortunately, the number went in the wrong direction. Again, the share of employees reporting a positive six-month business outlook dropped to 44.3% in February, which is down from 45.9% the month before. So right now, less than half of American workers think their company’s outlook is positive. Don’t like seeing that. And if you’re out in the job market and you realize that hiring is slowed, same thing is going to happen. The biggest year-over-year drop of any industry was information technology. That’s down 7.1 points. A
nd that’s despite what we know are massive AI and data center investments. The lowest confidence number overall from any industry was restaurants and food service. They were at the bottom, joined closely by manufacturing and government. Now, when you look at employee levels, the seniority breakdown is pretty telling. Senior level employee confidence has fallen over four points over the past year, which is the biggest decline of any group. And that’s been the case. That’s been a trend since the economic policy uncertainty spiked around May of last year. And of course, that has continued. So when senior level employees are losing confidence the fastest, that’s a signal that should catch everyone’s attention because these are the people closest to budget decisions, strategic planning, and restructuring conversations.
Let’s face it, they have more visibility into what’s coming than anyone else in the organization. And they’re the most worried. So that’s not a coincidence right now. So this confidence overall, if you’re an employer, when it’s below 50%, that should be treated as a retention warning. It’s not good because when your people, um, even if they’re not leaving today, it shows that they’re disengaged and in disengaged workers, they’ll underperform. So whether they’re leaving or staying for the wrong reason, that’s not a good thing. So if your business is healthy, make sure your employees know. If it’s not, but you have plans to do something about it, well, make sure they know that too, because confidence is tied very closely, as I said, to productivity, overall outlook, well-being. There’s so many reasons to make sure that employees are feeling good about where they are. And right now, this report shows that they’re clearly not.
ADP Weekly Pulse: Steady growth, but still carried by two sectors
When our final story today, ADP released its weekly jobs pulse this morning, and the trend is pointing in the right direction. ADP’s numbers have been more positive than others. They measure private employers specifically, and they showed that an average of 15,500 jobs per week have been added for the previous four weeks, and that’s holding steady after a number of weeks of strengthening. So what happened is December was low, right? December, the four-week average then was just 2,500 jobs, but a lot of that is seasonal. We know that. So it dipped and then it started to come back, and we’ve seen that climb. But look, that that’s not anything to get too excited about. If you add that up over a month, it’s still not great numbers, it’s still not where we need to be.
Um, I shared last week that their monthly number came in at 63,000. So we’re pretty much right on track there. That was a February number. But keep in mind that it these numbers are really concentrated in just a couple areas uh health services and construction. And the reason why the government’s numbers dipped last week were in large part due to that. Healthcare saw strikes with physicians’ offices and then construction, hiring was delayed by bad weather. That is true. Uh that that is happening. But I expect to see those bounce back. But again, we just have a couple industries carrying the entire market right now, and that certainly is not a good thing for the economy overall, for the workforce overall. So, you know, it’s a it’s an economy, or at least a job market that’s just stagnant right now. So there’s a couple of bright spots, that’s not enough, but that is where we are. So that’s it for today.
Fun fact: The Hawthorne Effect and what it says about being watched at work
Before we go, here is your fun fact: the Hawthorne effect. The Hawthorne effect shows that people work harder simply because they know they’re being watched. I’m being watched right now, sort of, by one of my dogs behind me. I don’t know if that makes me work harder. I don’t know if that counts for the Hawthorne effect, but sounds like micromanagement to me. That yes, people will work harder, but they probably won’t enjoy doing it. So pick your poison, I guess. That is where I will say goodbye. I’ll leave you with that thought. Thanks for listening. Please like, subscribe, share with anyone who you think might be interested, and I will look forward to talking to you tomorrow.
