Change management is hard, and it's harder when it comes to AI. Most business owners think you can give somebody the tool and they'll use it effectively.
We've seen it at one of my other companies. Everyone got the tools. Most people used them to create documents and help plan. A few really ran off with it and went deep.
Those few started with things they didn't enjoy doing. They automated them, and that gave them the space to be curious.
One of those problems was simple. Every Monday morning, somebody updated our scorecard with how our agents performed over the weekend. Logging each agent for each brand took 10 to 15 minutes. Now a person spends 2 minutes double checking the data and the formulas.
That's 8 to 13 minutes saved. It adds up every Monday, forever - roughly 7 to 11 hours a year from one spreadsheet. Find enough small things like that and you build a bank of time you get back. That's where the ROI comes from.
It's also hard to measure. Right now most of it runs on a feeling and on what people show off to the team. We're not counting agents. We're asking if the work is valuable, high quality, fast... and doesn't read like AI slop.
When a team treats AI as the solution instead of another tool, they build things they don't need, or they turn out bad work.
So here's what change management looks like. Work with your team to pick what to put AI behind first. Start basic, with something that doesn't need a lot of judgment. Have an open conversation, because some of your people think they'll lose their jobs over this. Tell them the tool is here to help, give them the space, name a few things you want accomplished, and check the quality at the end.
It always takes longer than we think. Once there's momentum, the rate of change picks up.
If you're not seeing results, it may not be the tool. It could be a change management problem. So ask yourself... am I actually seeing ROI from my AI?
I am going deeper on this live on Thursday, October 22 at 12pm CT. Save Your Seat ⟶
Visier, a workforce analytics company, surveyed 1,000 US full-time employees. 48% said they have exaggerated their AI use or expertise to colleagues or leadership. 45% feel pressure to use AI at work even when they aren't confident using it. Only 21% strongly agree their employer gives them the training and support they need, and just 28% think their leaders understand how employees actually use AI. Visier's research lead, Andrea Derler, warned that "leaders can mistake visible activity for meaningful transformation."
Our Take: If you judge adoption by who talks about AI most in meetings, you're measuring the performance. Ask to see one finished piece of work and how long that task used to take. Then ask what they'd hand off next if they had help.
Source: Visier
The European Central Bank asked about 20,000 people in 11 countries about AI at work. The share using it on the job went from 26% in 2024 to 41% in 2025 and 52% in 2026. The median user reports saving three hours a week, about 7.7% of their working time. Only about half of all workers both use AI and save time with it, so the gain across the whole workforce is closer to 3.8%. The authors add that saved hours only count "if workers turn the freed hours into extra output."
Our Take: Three hours a week is real, and it's easy to lose. If you don't decide where that time goes, your team will spend it on email and meetings. When someone on your team saves an hour, agree with them on what that hour is for.
Source: ECB Blog
Walmart wants every one of its 2.1 million employees to have some level of AI skills over the next few years. It started with role-specific AI certifications, built with OpenAI and Google, open to its 1.7 million associates in the US and Canada through its associate app. Chief People Officer Donna Morris: "We need to bring people along on the journey, and we need to make sure that we're educating people." Morris also said Walmart has about the same number of employees it had six years ago, with much higher revenue, and hasn't seen roles displaced by its AI use.
Our Take: Walmart trained people for the job they already do. You don't need a certification program to copy that. Pick one role, find the task that role likes least, and show that person how to hand it off.
Source: CIO Dive
In April 2025, Shopify CEO Tobi Lütke told employees that "reflexive AI usage is now a baseline expectation." Teams had to show why AI couldn't do the job before asking for more headcount. On The Knowledge Project podcast this month, he described what happens when that goes wrong: "We call those 'slop grenades' that people toss at each other. That's definitely a bad thing." He explained how they get made: "You don't really read it, and now it has to be reviewed by your colleagues."
Our Take: A mandate gets you more AI output, and somebody has to read all of it. Make that the person who made it. On your team, the rule can be that you read your own AI work before a coworker has to.
Source: Fortune
Gallup tracked about 30,000 worker responses from 2023 through early 2026. Roughly 3% to 4% of workers say their job is very likely to be eliminated within five years. Among people who use AI daily or several times a week, that worry runs at more than double the rate (6.3% vs. 3.1% in early 2026). Workers who feel their employer respects them and cares about their wellbeing are noticeably less worried, even when they use AI heavily. Workers who do fear for their jobs are less engaged and more likely to be looking for another one. Gallup's conclusion: "Better management may not erase workers' fears … but it can substantially ease those concerns."
Our Take: Gallup's data points at the manager. Before you roll anything out, tell your team what AI is for in your business and what you won't use it for. People who hear that from you worry less, and they're less likely to start looking elsewhere.
Source: Gallup
Ipsos surveyed 750 small-business owners and operators for the US Chamber of Commerce Foundation. 54% say AI has mostly helped with how long tasks take, and only 3% to 4% say it has mostly hurt. When AI frees up an employee's time, owners say 65% of them put it into more or better work, 56% into learning, planning or reviewing, and 36% use it to avoid overtime. Seven in ten owners say AI has changed what they expect from their employees.
Our Take: Most owners see the saved time going back into the work, which is the bank of time Brian wrote about. Write down one task you've handed off and the minutes it saves each week, then check the number again in a month and add it up.
That’s it for today!
— Brian, Murphy, John, Beth, and Jake. Some of the humans behind AI Front Desk.