The way we’ve traditionally approached compliance isn’t going to work particularly well for where we’re headed (see my previous post). Yes, we still have to know the rules, translate them into policies and processes, train people, document what happened, and respond when something changes. None of that is going away. But the rules are changing at the same time we’re changing work. And we have more choices about what work becomes than we’ve ever had before.
What work will people continue to do? What will AI do? Which jobs will disappear, and which will simply become very different jobs? What happens to performance expectations when technology allows one person to accomplish what used to require a team? What does management look like when technology is monitoring, analyzing, recommending, and sometimes making decisions about the people being managed?
We’re answering those questions now, whether we realize it or not.
McKinsey found a pretty significant gap between how much leaders thought employees were already using generative AI and what employees reported actually doing. C-suite leaders estimated the number of employees using AI for at least 30% of their daily work to be around 4%, but employee reports show the actual number is 3x higher.
We can intentionally redesign work, or work can be redesigned through thousands of individual decisions about how people use technology every day. Either way, we’re making choices.
What are we actually trying to accomplish?
We’ve spent decades using technology to make work more efficient, and AI promises to accelerate that enormously. Efficiency and productivity matter. Organizations need to grow, control costs, serve customers, compete, and make money. But I think we sometimes jump from “this technology can make us more efficient” to “therefore, efficiency is what we’re trying to accomplish.”
Those aren’t necessarily the same thing.
BCG’s recent research illustrates the opportunity, and the problem. Among frontline employees who regularly use AI, 42% said it was saving them at least a full workday each week. But 66% said they receive limited or no guidance about what to do with the time they’re saving, and more than half aren’t redirecting it into more strategic work.
That’s fascinating/terrifying to me. We’ve spent an enormous amount of time talking about how AI can create capacity. And yet, we’ve spent considerably less time deciding what we want to do with it.
We’re creating capacity. But capacity for what?
If AI allows one person to do work that previously required three people, we might grow without adding headcount. We might eliminate two positions. We might ask that one person to produce three times as much. We might redesign the job entirely. We might stop doing work that never added much value in the first place. Or we might use some of that capacity for the judgment, relationships, creativity, development, and customer interaction that continually get squeezed out because there isn’t enough time.
Those are very different outcomes from the same technological capability.
The important thing is recognizing that we’re making a decision.
That phrase and this feeling is familiar to me because I used to say some version of it in my alternate life when I taught the course American Constitutional Development at Baylor University.
I taught during the least coveted Tuesday/Thursday 3:30–4:50 time slot. My classes were largely senior music students whose practice schedules conflicted with almost everything else, seniors who had put off the requirement as long as possible, fraternity gentlemen who explained that later classes better accommodated their sleep and nap schedules (their statement, not my assumption), and sophomores who simply didn’t know better yet.
It was a tough crowd.
The purpose of the course was to help students understand the foundations of American government and how they had evolved, largely through court cases, so they could make informed decisions about their own role as citizens. When we talked about voting, I wasn’t interested in telling them who they should vote for (though I have strong opinions); I wanted them to understand why they were making the choice.
If you vote for someone because you agree with their view on particular topic, great. If you vote for someone because you like their smile, that’s your choice too. Either way, be intentional. Have the integrity and courage of your decision. Don’t pretend you didn’t know what you were choosing. Don’t act later as though you didn’t have a choice.
Recognize that you’re making a decision.
I find myself thinking about that a lot as we talk about AI and the future of work.
If we eliminate positions to improve margins, we should be clear that we’re choosing to use the productivity gain that way. If we use AI-generated capacity to increase output expectations, that’s a choice too. So is redesigning jobs to give people more time for judgment? Relationships? Creativity? Customers? Or something else?
Some decisions will be good for the business and hard on employees; there will be tradeoffs. But we should know what we’re optimizing for, what we’re trading away, and what the consequences are likely to be for the people affected.
Human outcomes belong in this conversation because they don’t stay neatly in the HR column. They show up in trust, turnover, productivity, quality, customer experience, risk, and the organization’s ability to sustain whatever new way of working we’ve created.
My friends Jason Averbook and Jess Von Bank deep dive in these waters in The Edge of Now podcast.
Then there are the rules.
This is where strategic compliance comes in
Employment law has developed, in part, around places where the relationship between business interests and consequences for people created enough tension that governments, courts, or regulators stepped in. Workplace safety, wages and hours, discrimination, disability, leave, privacy, worker classification - none of those became regulatory issues in a vacuum.
We’re creating new tensions now.
AI is changing what employers can know about employees, how closely work can be monitored, how quickly decisions can be made, and how much human judgment is involved. We’re reconsidering who performs work, how jobs are structured, and what the employer-employee relationship looks like.
Meanwhile, regulators, legislatures, and courts are working through many of the same questions.
The absence of an “AI law” doesn’t mean the absence of law. Existing discrimination, disability, privacy, wage and hour, and other employment laws still apply. Let me say that again just to make sure you hear it above the political/marketing noise: laws still apply.
At the same time, new laws and proposals are emerging as governments decide where the existing rules aren’t enough and where they stifle innovation. We don’t know exactly where all of that will land, which is why strategic compliance requires more than monitoring what changed.
Regulatory intelligence looks at what is happening across legislation, litigation, enforcement, regulation, and jurisdictions and asks what those signals might mean for the business: Which workforce decisions could they affect? Where might current practices become vulnerable? What patterns are emerging? And who needs that information before the organization makes its next move?
One proposed bill, one court case, or one new state requirement doesn’t determine destiny. But when the same concerns start showing up in different places, that’s useful information. While we don’t know exactly where the boundary will eventually be drawn, we can often see where people are starting to argue about where it belongs. We need to pay attention to that before we decide what to build.
Consider performance management, for example. AI can analyze work, identify patterns, surface performance concerns, recommend coaching, identify high performers, and give managers information they could never realistically collect on their own. There could be enormous value in that.
There are also a lot of decisions buried inside it. What data are we collecting? What is the system inferring? What happens when someone’s performance pattern is affected by disability, leave, or an accommodation? Are the measures we’re using creating different outcomes for different groups? Are the measures we’re using actually impactful? What happens when the system gets it wrong? How much discretion does the manager retain? Who ultimately owns the decision?
Those questions aren’t reasons to abandon the technology. They are things worth knowing before we’ve selected the system, signed the contract, built the workflow, trained the managers, and evaluated thousands of employees through it.
That’s the difference I’m trying to get at with strategic compliance.
Sometimes regulatory intelligence changes what we do. Sometimes it changes how we do it. Sometimes the business makes exactly the same decision it would have made anyway, but with a clearer understanding of the regulatory implications, business tradeoffs, and consequences for the people involved - and that’s valuable too.
Before we decide what work becomes
We can’t predict what employment law will look like five years from now, and we shouldn’t try to design work around regulations that haven’t been written yet. But we can pay attention.
Maybe the more useful question is: What are we building today that could become very expensive to change tomorrow?
The answer is worth taking a beat to consider. Maybe that changes the decision and maybe not. Maybe it simply means making sure a vendor contract gives us enough visibility into how decisions are made, or retaining the ability to change what data a system uses, or auditing outcomes, or making sure we can change a workflow without rebuilding the whole thing.
Because I can absolutely imagine organizations spending the next two years redesigning work, investing in technology, changing operating models, training everyone, working through the inevitable problems and finally getting really good at a new way of working right about the time they discover they can’t continue doing it the way they designed it.
That’s an expensive time to start thinking about compliance.
I’ve argued that the law gives us the floor, and organizations decide what they build above it. We’re about to do a lot of building.
Strategic compliance gives us a way to bring regulatory intelligence into those decisions while there are still choices to make, alongside what the business is trying to accomplish and what those choices will mean for the people affected by them. It doesn’t make the decision for us.
The whole point is that we’re still responsible for the decision.
AI is expanding what’s possible. Before we decide what work becomes, we should understand what we’re building, why we’re building it, the rules we’re building within, where those rules may be headed, what our choices will mean for the people who will actually live with them, and how to prepare for doing business under them.
Then have the integrity and courage to own what we chose.