Artificial intelligence is no longer something businesses can afford to watch from the sidelines. It's already inside your organization, and your employees are using it whether you've sanctioned it or not.

What separates businesses that benefit from AI and those that don't usually comes down to one thing: whether leadership gets involved early enough to shape how it's used.

Jimmy Hatzell, CEO and co-founder of Hatz AI, has spent years at the intersection of cybersecurity and enterprise technology. He sat down with Scott Kreisberg, host of the One Step Beyond Cyber podcast, for a candid conversation about what AI adoption actually looks like inside real businesses.

What he describes is messier than the commercials suggest, and a lot more transformative than most leaders expect. What follows draws directly from that conversation, with insights business leaders can put to use immediately.

 

The "Set It and Forget It" Fantasy Doesn't Hold Up

There's a persistent myth about AI, that you hand off your work, step away, and come back to a finished product. One Super Bowl ad even suggested as much. Jimmy addresses it directly:

"I think there's even a Super Bowl commercial about this. Give your work to AI and then go have a barbecue. And it's just not the reality of what happens. What you get instead is 10x the output. And then it's a different problem where you can have a junior employee producing 10x the output that a senior employee needs to review."

That reframing matters because AI doesn't eliminate work; it multiplies it. A junior analyst can now produce a volume of reports, research, or drafts that previously would have required a full team. The bottleneck shifts from production to review, from doing to evaluating. For business leaders, that means the role of experienced employees evolves rather than disappears. Senior staff become reviewers, editors, and decision-makers at a higher rate than before.

This is a fundamentally different operating model, and companies that understand it early are the ones positioning themselves well.

 

AI Adoption Is Coming From the Bottom Up, and That's a Problem If You're Not Ready

Every major technology shift in the last 30 years, moving to the cloud, adopting new security software, rolling out Enterprise Resource Planning (ERP) systems, was driven from the top down. Leadership identified the need, IT implemented the solution, and employees were trained to follow along.

AI is different.

"It's the first technology movement that I've seen or that I've really heard people talk about that's been bottom up instead of top down," Jimmy explains. "Nobody was ever like, 'I really prefer this antivirus, I signed up for it without anyone knowing and installed it on my machine.'"

But that's exactly what's happening with AI. Employees are already sharing company data with ChatGPT, Claude, Grok, and a dozen other tools while using personal accounts. A sales rep reconciles a spreadsheet faster by uploading it to a free AI tool. A marketer drafts client emails in a personal AI account. A project manager uses a browser extension with AI built in.

The risk isn't theoretical. If an employee uploads a company financial document to a personal ChatGPT account every month for reconciliation, that data may be used to train the model, or worse, accessible in ways that violate your confidentiality agreements. When someone leaves the company, they may take two years of operational insights with them without anyone knowing.

The solution isn't to lock AI down. That approach just pushes usage further underground. The solution is to provide a secure, managed environment where employees can use AI without creating compliance or data exposure risks.

 

Blocking AI Creates Shadow IT. Secure Access Prevents It.

One of the most common mistakes organizations make is treating AI the way some companies treated social media a decade ago, blocking it entirely and hoping the problem goes away. It doesn't.

"Without adopting, without these controls in place, what you're left with is people doing it anyway," Jimmy says. "People uploading things to their personal accounts, all sorts of problems from that, whether it's training on the data or someone leaving the company with two years' worth of financial information."

Shadow IT, employees using unsanctioned technology, is an old problem. But AI has accelerated it dramatically. The better path is what security professionals call "responsible AI," giving employees access to vetted, secure tools, establishing clear data handling policies, and building a culture where people know what's allowed and why.

That means:

  • Identifying and approving specific AI tools for company use

  • Establishing a clear data classification policy (what can and cannot be shared with AI)

  • Training employees on appropriate use, not just prohibitions

  • Creating visibility into how AI is being used across the organization, without invasive monitoring

"The best way is giving people a secure and safe place to use AI," Jimmy notes, "and just guiding principles, a bit of training and guidance and access to tools, to just let them figure things out on their own."

 

Knowledge Workers Are Facing a Real Shift, and It's Coming Fast

Software engineering has arguably changed more in the last two months than it did in the previous two decades. Engineers at major technology companies are spending less time writing code line by line and more time reviewing, directing, and orchestrating what AI produces. The same shift is headed toward every knowledge-worker role.

Think about the people in your organization who build reports, analyze financials, draft contracts, or handle complex documentation. Those tasks are increasingly within AI's capability. According to Jimmy, the near-term impact is significant:

"I think we're going to see it really hit knowledge workers over the next year or two. Where people will be able to reconcile those reports, run the discounted cash flow, do a P&L report, things where they generally know what it is, but they haven't done the specifics before. And there will be a lack of expertise when the person is actually reviewing the work."

This creates a specific leadership challenge. When AI narrows the gap between what a junior employee can produce and what a senior employee has spent years learning, the review process becomes more demanding, not less. A CFO who previously managed six direct reports may suddenly find themselves reviewing far more work and with less confidence that the underlying process was followed correctly.

The answer isn't to slow AI adoption. It's to build your team's ability to work alongside it and to evaluate AI output critically.

 

How to Roll Out AI Across Your Business

If you're leading a company with anywhere from 50 to 500 employees and you want to implement AI in a way that actually sticks, Jimmy's framework is practical and direct.

The approach is both top-down and bottom-up simultaneously.

Top-down means setting guardrails:

  • Define your data classification policy, and what can and cannot be shared with an AI tool

  • Identify and approve the tools your team is allowed to use

  • Set disclosure expectations (when does someone need to indicate that AI contributed to a deliverable?)

  • Roll out training that's accessible and relevant, not just a compliance checkbox

Bottom-up means giving employees room to experiment:

  • Don't limit AI access to a small pilot group or a senior leadership team

  • Avoid mandating specific use cases from the top; let employees find what's useful in their own work

  • Create a shared space (a Teams channel or a Slack group), where people can share what they've discovered

  • Celebrate wins. Who found the most useful AI workflow this month? Who saved the most time?

"You give AI to all the employees, and a little bit of training and guidance on what's allowed, maybe some examples of how to get started, and they'll get to the same place, and it'll be a lot simpler and often better, because they're the ones actually clicking the buttons and doing the things."

The companies that see the fastest, most sustainable AI adoption are the ones that stop treating it as an IT rollout and start treating it as a cultural shift.

 

Replacing Employees Is the Wrong Strategy

There's a tempting but shortsighted line of thinking that AI adoption means reducing headcount. Some companies are pursuing it. Jimmy is direct about why it's a mistake, not just ethically, but operationally.

"I think it's a mistake for companies who are like, 'We're going to fire all our employees and replace with AI.' I don't think that's a good idea. I don't think it's the right thing to do for people, but I also just don't think it's a good thing for business. Our goal should be to upskill employees and help them in an AI world."

Here's the practical case. The institutional knowledge, client relationships, judgment calls, and contextual understanding your experienced employees carry cannot be replicated by a language model. What AI can do is handle the repetitive, time-consuming tasks that eat up their capacity, freeing them to focus on the work that actually requires human judgment.

An employee who previously spent four hours a week on manual reporting and now does it in 30 minutes with AI isn't a liability. They're a significantly more productive asset. Replacing that person means losing everything they know and starting the upskilling process over with someone new.

The smarter investment is clear, help your current team get fluent with AI tools, remove friction from that process, and measure the gains.

 

Practical Steps for CEOs Who Want to Act Now

Jimmy's recommendations for business leaders who want to make progress in the next quarter, without creating new risk:

1. Use AI yourself.

Not just for curiosity. Build a real habit. Use it for meeting prep, reviewing contracts, drafting communications, and summarizing reports. Leaders who talk about AI without actually using it daily lose credibility with teams who are already relying on it. You don't need to be technical; you need to be curious.

2. Give your whole team secure access.

Not just a pilot group. Not just the VP and a few department heads. Everyone. Set the guardrails first, data policy, approved tools, basic training, then let people explore. You'll be surprised by what surfaces.

3. Create a culture of sharing, not fear.

Many employees are already using AI for their work and haven't told anyone. Some are worried about how it will be perceived. Create explicit space to share wins. Recognize the team member who found a clever use case. Make it clear that using AI to work smarter is something to be celebrated, not hidden.

As Jimmy puts it, "You can't make your team feel like they're using AI to remove themselves from a job. It's, we're upskilling together, and we're learning together to make us all more skilled employees."

One practical tip on where to start: connect AI to the tools your team already uses every day. Email and calendar integrations tend to drive fast adoption because the value is immediately obvious. If you have a sales team, connecting AI to your CRM removes one of the most-hated tasks in sales, updating activity notes, and suddenly the data quality improves dramatically.

 

The Security Question Deserves a Straight Answer

Because so much of AI development is happening fast and in experimental environments, the security landscape is genuinely uneven. Not every AI tool has been built with enterprise security in mind. Plugins, integrations, and third-party AI connectors vary widely in how they handle your data.

Jimmy's background in cybersecurity informs a clear position: the answer to security concerns is not to avoid AI, but to vet carefully.

Responsible AI isn't about caution to the point of inaction. It's about building a structure where employees can be bold and productive without the organization carrying unnecessary risk.

 

Where This Is Headed

Model providers estimate that by the end of 2027, AI will reach a point of recursive self-improvement, where models begin identifying gaps in their own performance and training themselves to get better. That's a significant threshold, and it's closer than most people realize.

Between now and then, the competitive advantage goes to organizations that build what Jimmy calls "AI muscles," teams that are genuinely fluent in working alongside AI, who default to it as a tool, and who know how to evaluate its output critically and push it toward better results.

"Building your AI muscles is really the important thing," he says. "Pushing the boundaries of what AI can do, understanding what's actually possible, that's how you prepare."

The businesses that invest in that fluency now will operate faster, produce more, and adapt more quickly than those that wait.

At One Step Secure IT, we help small and mid-sized businesses adopt AI in a way that's practical, secure, and sustainable. Whether you're just getting started or looking to build out a more structured AI strategy for your team, we're here to help.

Check out the full podcast episode in video format on the One Step Secure IT YouTube channel!

Schedule a conversation with our team to find out what responsible, effective AI adoption looks like for your business.