Xennials Will Navigate Us Through AI

We lived through an analog childhood and a digital adolescence. That turns out to be the best preparation for the AI moment anyone could have designed on purpose, and it starts with a printout.

If you learned to drive before smartphones, you remember the MapQuest printout. You typed in two addresses, printed four pages of turn-by-turn directions, and left them on the passenger seat. You also remember the moment, usually somewhere around step 11, when the printout told you to turn left onto a road that had been closed for years and you had to work it out from what you actually knew about the city.

That moment turned out to be the most useful AI training I ever got. The printout was right most of the time, it was fast, and it sounded exactly as confident when it was wrong. The skill was never reading the directions, it was knowing the road well enough to notice when the directions stopped matching it. Then GPS started talking us through every turn, and the people who'd learned the city first were still the ones who noticed when Apple Maps tried to route them into a lake.

Xennials, the micro-generation born roughly between 1977 and 1983, were the last kids to grow up offline and the first teenagers to log on. Gen X already claimed the name MTV Generation, so we need one of our own, and I'd like to nominate the AIM Generation.

We were the first to hang out with our friends online after school, the first to agonize over an away message, and the first to learn that the tone of a typed sentence can start or end a friendship. If you're a Xennial, you can already hear the door creak of a friend signing on. If you're not, AIM was AOL Instant Messenger, the first place any of us slid into someone's direct messages.

All those hours at the family computer taught us something. We didn't adopt the internet as adults, and we didn't grow up assuming it had always been there. We watched it arrive, then spent school and our early careers doing every job twice, once the old way and once the new way, while a teacher or a boss decided which one counted. I clashed with those people constantly, especially in high school, where I kept pushing new ways to format, design, and write things before anyone had decided they were allowed.

Research is the clearest example. We started with a set of World Book encyclopedias at home and a library card for everything they didn't cover. Then came Encarta on CD-ROM, which felt like the future and was out of date the moment it shipped, exactly like the books on the shelf. Then the internet made information current, and Wikipedia kept it current by letting anyone edit it, which meant we traded stale facts for facts that needed checking. Every step fixed the last problem and introduced a new one, and the skill that survived all of them was knowing where a fact came from before you trust it. AI is the next step on that path. It's current and fluent, and it doesn't always know where its facts came from either.

That gives us a specific kind of literacy that matters right now. We're comfortable enough with the tools to use them all day, and we remember enough of the before to know what the work actually is underneath them. When an AI draft comes back polished and wrong, we notice for the same reason we noticed the closed road, because we knew where we were going before the machine told us how to get there.

We've also heard this one before. In 1995, right as everyone was calling it the information superhighway, Jeremy Rifkin's The End of Work predicted that computers would wipe out tens of millions of jobs and hollow out the middle class. What actually happened is that the work changed shape, and the people who understood both the old way and the new way ended up running it. The AI version of that prediction is louder, but it isn't new to us. If anything, the usage caps on today's AI tools feel a lot like AOL charging us by the hour to be online, which is a good reminder of how early we still are.

The HR Certification Institute, known as HRCI, published its first State of HR report this year, a survey of more than 4,500 HR professionals, and it makes a similar point from the top of the profession. Amy Dufrane, HRCI's chief executive officer, put it this way. "Our generation has already seen massive changes in technology, so it is not surprising that many companies are already leaning into AI." What surprises her is how few of them are training their HR teams to use it.

That isn't a claim that other generations can't do this, and plenty of people of every age are doing it well. Xennials have just had more practice than anyone at standing with one foot on each side of a technology shift, and the AI moment rewards exactly that posture. Use the tool hard, trust it conditionally, and never lose your own sense of direction.

If you're not a Xennial, you can still pick up these habits. The first habit is knowing the shape of the answer before you ask. Sketch your own outline, your own numbers, or your own first take before the AI hands you one, so you have something to check its draft against, and ask where every fact came from.

The second habit is doing the job both ways. Run the old way and the new way side by side for a few weeks until you know where the new one breaks, then hand it the parts it's good at. We didn't wait to feel like experts before we logged on, and nobody should wait for that now.

The fastest way to pick up either habit is to sit next to someone who already has it. Pair people across generations on real work instead of a lunch-and-learn, and let the learning run in both directions, because the Xennial in that pair has plenty to pick up too.

That pairing only works if both people show up. A Xennial friend of mine was recently matched with a longtime AI detractor from an older generation for a session the company required. The detractor joined the call, announced they weren't interested, and left.

Xennials will help you, get you across the finish line, pick you up off the floor, and stay with you until it clicks. We also move fast and run low on patience, and we will gladly leave you behind if you don't take the hand we're holding out.

I run my day-to-day work in HR with a small set of AI agents, each with its own lane. One handles performance program operations, one handles compensation job matching, one handles workforce data, and one works as an executive assistant that holds my calendar, the people context, and the follow-up list across all of them.

The rules matter more than the agents. The rule I'd give anyone is that sensitive information, like someone's pay or a personal situation, never gets passed between tools on autopilot. Every handoff between agents is a written document that I read before it goes anywhere. Anything that still needs a send or a deadline from me stays on the assistant's list, because a handoff is not a close.

That setup can get me through a morning of manager workshops, follow-ups, and a full scorecard audit before lunch. The week of meeting prep that used to come before a stretch like that now fits into an afternoon. The speed is real, and it's only half of what changed.

The agents handle the turn-by-turn, but I still choose the destination, and the standard I aim for is checking everything against what I know is true before it goes out.

Whether you call it Human Resources or People Operations, it gets treated as a cost center because our work is felt but not shown. When onboarding goes well, nobody sees the forty things that went right. When a manager handles a hard conversation better because HR prepped them, the manager gets the credit, which is correct, and the People team stays invisible, which becomes a problem at budget time.

AI changes that, because the same tools that do the work can show the work. A structured brief, a pattern analysis, or a decision-ready note now exists where there used to be "I handled it."

It also changed who has to wait. Xennials grew up figuring things out on our own, usually because nobody else in the house could program the VCR, and that resourcefulness has never been more useful. Before AI, an HR dashboard meant weeks of asking a data analyst for help, and People always sat at the bottom of their queue because we don't generate revenue directly and we don't face customers. By the time the insights arrived they were already lagging, if they arrived at all. Tools like Claude cut out that middle step, so the person who understands the question can now build the answer.

The clearest example is a pair of dashboards. One is a talent dashboard built on the performance program that shows the talent bench and turns the scores into talent analysis with themes leaders can act on, on top of the basics like completion and how individual managers tend to score. The biggest theme it surfaces is manager effectiveness, which is arguably the strongest driver of change a company has, if it's willing to invest in moving that needle. The other covers the workforce (headcount, trailing-twelve-month attrition split voluntary and involuntary, tenure, span of control, engagement). The design choices mattered more than the charts. Every panel names its population, its date, and its formula, and numbers like tenure show both the median and the average. A toggle separates employees from contractors, because every HR practitioner has sat in a meeting where two people quoted different headcounts and both were technically right. No individual names appear anywhere, and compensation panels stay off until the job matches underneath them are verified, because a chart built on inherited data is a chart built on sand.

Any good executive will tell you a snapshot without a timeframe is meaningless, so build to that standard from day one. When your executive team goes looking for the seam in your numbers and can't find one, that's what visibility buys you.

Neither of these needs a budget or a vendor.

Knowing the road is the job now, because the tools will keep getting better at directions and the People leaders who stay valuable will be the ones who still know where they're going and can show their work on the way there.

Almost every founder I talk to says AI fluency matters, and it's all over job descriptions now. What's missing is what happens after the offer. Most companies haven't defined what good looks like in the role, built it into performance expectations, or covered it in onboarding, so every employee is left to decide alone how much to use it and how far to trust it.

HR is in the same spot. In that HRCI survey, 53% said their company offers no AI training at all, and most of what does exist is homegrown, mostly published materials, in-house classes, and lunch-and-learns. Only 8% had been offered any kind of AI credential, which means the function expected to roll AI fluency out to everyone else is mostly teaching itself.

The fix is to treat it like any other core competency, with explicit expectations at every level from individual contributor to exec. Define what good looks like in each role, then give people room to experiment inside clear boundaries, the same way the best managers handled every new tool we've ever been handed.