Many of my conversations about AI start with loss. People bring up large companies reducing their teams, with some openly saying that AI and other tools are taking on work people used to do. They talk about organizations with a lot of senior, high-salaried roles looking at early-career hires who use AI tools to produce similar output. Roles in tech, marketing and consulting are being cut or reduced, and buyers are questioning what work is worth when it used to cost more.

For the people and families going through it, it isn't a narrative. It's their lived experience. Those aren't my organizations, and I don't have insight into what's happening inside them, so it wouldn't be fair for me to comment on anyone's job or a company's decisions.

This is easier for me than for many others. I have years of experience to fall back on, and I work for myself, so I can carve out time to learn. It's harder for those earlier in their careers and for anyone already affected.

What I can share is what I hear. As a consultant who works with AI tools every day, I've watched them change my own work, and that keeps me adapting to where the work and the labor market are going. In conversations across business, I hear from three groups: people who use AI for everything, people who dabble, and people who are afraid of it. These are conversations, not research, and each of them has shaped how I think about where AI helps and where people still do the thinking.

I listen to where people are coming from before sharing a view on AI

These conversations happen naturally, with people working in all different areas of business. I try to understand where each person is in their career, whether that's early, mid or late stage, how much time and appetite they have, and what they've heard and read, along with where those sources come from.

Listening comes first. I want to understand what's driving someone to think one way or another before I say anything. Where it makes sense, I try to be helpful. Other times, it's a journey that person needs to go through, and it isn't my place to tell them what they should be doing because of where their market might go.

Seasoned experts who dabble with AI often say their business moves too slowly

Some of the people who dabble with AI are seasoned business professionals with deep expertise in their fields. When I suggest they figure out how these tools fit into their workflow, the answer I often get is some version of "No, my clients move too slowly, and their industries won't be affected. I don't need to change how I work."

My reply is that it isn't about how fast their clients move. It's about staying current with where the technology and people's interest are going, especially in consulting. AI tools can now help someone find, organize and work through much of what's been written in their field.

That makes a subject matter expert who uses these tools more valuable, not less. Someone with a deep understanding of the subject can work through information faster, judge whether an output is right, and see how it applies to the situation at hand. Without that depth, it's easier to trust whatever the tool gives you.

The heavy AI users I talk with rarely mention what they do with the time saved

Then there's the opposite group. Some of the heavy users I talk with are deeply invested in using AI across every part of how they work, from operations and workflow through to what they're paid to do. A few aren't doing much of that themselves anymore, or overseeing it closely.

The caution I share is about losing touch with what you're doing and why you're doing it. The more you automate, the more the thinking moves over to a system. If you aren't practicing yourself, you lose some sharpness. A tool that started out supporting your expertise ends up replacing it, and when the systems grow and change, you may not be able to adapt, or even hold a conversation about the subject.

That caution applies to me too. Early in my agency career, I learned never to be above the work. Even in senior leadership roles, I kept enough hands-on ability to do it myself. Reading and watching content about AI isn't enough to stay sharp, so I practice, read and experiment hands-on. Working for myself gives me the time and space to do that, which not everyone has.

In these conversations, the heavy users rarely bring up what they do with the time they gain back, or how these tools help them support the people they work with. That's the part I'd most like to hear more about.

Fear of AI often starts with seeing your expertise done in a few keystrokes

A conversation with a very seasoned business professional, closer to retirement than not, stays with me. Their fear is being automated out of their job before they reach retirement age. I'm empathetic to that. Depending on the person, picking up new skills on top of a full workday and a full life can be overwhelming, and finding the time and space to learn is hard.

My son has his own version of that fear. I check in with him often to hear what he thinks about AI, and one of our first conversations started with "AI slop." When I asked what that meant to him, he got specific. He's a big gamer, and what he'd read and watched online was about AI taking jobs in video game development. That's a real concern for many people in that industry, so I listened before sharing anything. Helping him question AI answers has been a bigger project for us.

Fear often comes from seeing a tool do something you took pride in. You built deep expertise and did it by hand, and now it seems to take only a few keystrokes. In my own design, copywriting and marketing work, AI has moved my time from producing every piece by hand to framing the problem, checking what comes back and judging it as a whole. AI is a bigger shift than the templates people have bought for years to speed up their work, but the same principle applies: someone still has to know what good looks like. Talking through specific scenarios together is what helps people see that.

An HR onboarding handbook shows where the time saved by AI can go

I've done HR work in the private sector, alongside my leadership roles at an agency and a software startup, and sat on policy committees in the public sector that met quarterly or annually. At the agency, I worked with a consultant to design and develop a full HR policy manual, end to end. At the software startup, I rebuilt one from what I knew. It went faster, but it still took a lot of time.

Writing was only part of that time. Each manual took a long time to write by hand, and then even more time for people to review it, understand it and roll it out.

That background came to mind during a consulting project. While organizing an internal committee, I suggested including someone from HR, since some of the research we were doing could help with their onboarding and training. The team was updating its handbooks, starting with new team member onboarding, and the first draft wasn't expected for several months.

From my own work on these documents, scheduling, getting the right people in the room and shifting priorities are what usually slow them down. With AI assisting and the right inputs, a new version could come together much faster, with confidence:

  • What already exists. The current documents, which in the past were reviewed by hand as the industry or the organization changed.
  • Secondary research. Best practices and where HR, or people and culture, is heading, kept free of anything proprietary.
  • Conversations with people. Talking with the people affected, both directly and indirectly, to understand the what, the how and the why of the change. This is the most important input of the three.

Those three could turn a draft that takes months into one that takes days, with more of the time saved going to reviewing, understanding and rolling out the change. My confidence comes from my own research work. Organizing transcripts from customer and employee discovery interviews and workshops used to mean going through every one by hand. Now I query them in a project to pull out and organize the insights, with no personal or identifying information. Having done the HR work manually and the research work with AI, I'm confident the same approach carries over.

Speed doesn't replace review. An HR handbook carries legal and compliance requirements, and accuracy matters. A balanced way to start is with one section rather than the whole handbook, with HR and whoever handles legal and compliance reviewing it before anything goes out.

I raised it with the COO as an example of where an AI-assisted process could speed things up. The goal isn't to replace the HR person who manages these documents. It's to support them with a new process, the same way other tools already support them. What an organization does with the time saved is its decision. My view is to put it back into people: less time writing the handbook, more time with the people you're writing it for.

The COO was curious and wanted another opinion, while busy running the business. We didn't take it past that conversation.

Handling sensitive information with AI takes the same care as before

Some people question putting transcripts or HR documents into AI tools at all, and it's a fair concern. Long before AI, transcription services, survey tools and cloud documents handled the same kind of material, and the same diligence applies now.

I work for myself, without a privacy team or a policy department, so my approach suits a small team just as well. I keep a clear line on what never goes into a tool. I use business accounts rather than personal ones, and check their data settings. Names and identifying details come out before anything goes in, and people give permission before anything is recorded.

When the material belongs to an organization I'm working with, its AI use policy comes first, and its privacy or legal team is the place to check anything sensitive. When in doubt, leave it out.

Judgment comes from knowing the subject, not from the AI output

The harder conversations are about relying on output you can't judge. When someone without deep knowledge of a subject uses AI to write about it, it can feel like an accomplishment, but what comes out doesn't sound like them, because they can't judge what the tool produced or give guidance on how to improve it. That can happen at any stage of a career.

Judging what AI writes for you takes two foundations: communication skills, and an understanding of the subject being written about. Without them, you're passing your judgment and your expertise to a tool you can't evaluate. It usually shows in conversation, where the depth isn't there yet.

The same thing comes up in budget conversations. It isn't uncommon to hear, "Why would I pay for this when I can just use AI?" My answer is to try it. For some tasks, the output will be good enough, and they won't need outside help. That's a fair outcome. For others, they'll see where expertise still matters. A price based on output alone misses the expertise behind it.

Judgment grows from working alongside people who have it. Someone starting out can learn a playbook. The harder part is learning to think when things are unclear, to read a room, and to know when to push back or ask for clarification. The tasks that look easiest to hand to a tool are often where newer team members are forming that judgment, so they're worth keeping close to the people they're learning from.

AI gets me to information I trust faster, and the people side stays with me

In my strategy and digital marketing work, I've always started from prior projects and what I know. I would also spend a lot of time on secondary research: listening to podcasts, and reading articles, websites and books. That research is still valuable. With my judgment applied, I now get to high-quality information I'm confident in much faster, and that saves time and adds value to every project.

What I don't hand over is the context: why a tool is being used, and how. Even when I share all my thinking with an AI tool, it doesn't capture the people side. That's where I spend most of my thinking. The advances in AI are exciting. What I focus on most is how they affect the day-to-day lives of the people I'm working with.

If anything, AI gives me more time for that part. I have more conversations, so my understanding is well rounded and I'm not making assumptions.

Experimenting with AI is the first step, even when time is short

If there's one thing I'd want a reader to do, it's experiment. Diving into a new tool can be intimidating, especially for people who are already stretched, and I've heard "I just don't have the time to learn this" more than once. I understand that. If you can find even a small window, it's worth it.

That window doesn't have to come from your evenings or downtime. Try it on a task already on your list, so the experiment happens inside time you were going to spend anyway. When I train teams on marketing with AI support, I suggest starting with something high value and low risk: one small task where you already know the answer, something that would take you longer to do by hand. Because you know what good looks like, you can judge what the tool gives you and learn where it helps and where your expertise still matters.

Ways to start using AI and keep people at the center

These move from the first conversation to putting it in place.

  1. Listen first. Understand where people are in their careers, what they've heard and read, and what's driving how they feel about AI.

  2. Talk through specific scenarios. Walk through how a tool would fit a task people already know, rather than what they've read in a headline.

  3. Start with one high-value, low-risk task. Choose something already on your list that you know the answer to, like one section of a handbook rather than the whole thing.

  4. Bring the right inputs. Start from what already exists, add secondary research, and talk with the people affected.

  5. Keep sensitive information out. Write down what never goes into a tool, and remove names and identifying details before anything does.

  6. Keep review with people who know the subject. Include legal and compliance review where the content calls for it.

  7. Keep practicing hands-on. The more you automate, the more deliberately you need to stay sharp.

  8. Decide what the time saved is for. Put it back into the people side, like onboarding the people the handbook was written for.

  9. Pair newer team members with experienced ones. Give them the tasks where judgment forms, alongside someone who can show what they look for.