It's hard to believe it, but it really has been two years since we launched Hatz AI in early 2024. AI has improved incredibly fast since then, changing how people expect to use it and how they approach their everyday work tools. As AI gets better at handling complex human tasks, it's even making people reconsider their career paths.  

At our Partner Roundtable this February, our CEO and Co-Founder Jimmy Hatzell shared his thoughts about these last two years with Hatz AI Adoption Manager Mark Leeper. They discussed why traditional software roadmaps have evolved into something more like high-frequency trading, the psychological shifts hitting the workforce in 2026, and why the "evaluation loop" is the most dangerous place for an MSP to be right now. 

The Conversation

Mark Leeper: Jimmy, I was looking at the calendar, and I noticed that it's about 2 years since Hatz officially launched. I'm curious to hear your thoughts about how the AI landscape has changed over the last two years, some of the big things that you've seen? 

Jimmy Hatzell: Yeah, I mean, what I've learned over the past 2 years is you can't predict AI longer than a couple weeks or months at a time on what's going to be possible, or what you could do. If you asked me two years ago at this time, I would have told you that by now, everyone would be fine-tuning their own AI models and running their own fine-tuning workloads, which is something that nobody really does for general use purposes. 

I don't think anyone knew that the models would get as good as they are today, and how important tool use has been. It's just been incredible how fast things have moved. In our world, what we do is curate the best LLMs, use cases, products, and services for AI, and make them available to our partners to manage and sell. 

My role as CEO and working with the product team has been more like a hedge fund trader than a traditional software engineer. So people ask me, "What's your roadmap for the next 12 months?" And it's like, we don't know where AI is gonna be at in 12 months. Nobody knows, right? We read the news, we try new products, we test things. Anthropic could release Claude 5 today, and that's what we'll be releasing today. It's been very interesting and dynamic, but it's been fun. 

Mark: It's interesting that you talk about the rate at which tools and things outside of AI have also had to keep up with the pace of how fast AI is moving. It's like everything else around it has also been lifted and had to evolve. People moving away from just traditional Google searches and legacy technology – that whole way of interacting with data has had to change, too. 

Jimmy: Yeah, and you know, Mark, I think a lot of people aren't ready for the human effects that are gonna hit in 2026. Now that AI has gotten very good, especially with tool use and longer-running tasks, you start to see a psychological thing that happens with people. 

You might work on getting really, really good at something for a very long period of time-maybe a 10, 20-year career. And then, all of a sudden, someone who didn't work that long is able to produce similar results to you using AI in a very short period of time. The psychological effects that happen to people when that starts to happen-we've never really dealt with it before. 

We've seen it in micro areas of business, like Chegg – their stock dropped 99%. And now, as AI is getting really good at doing a lot of different kinds of work, I think it's gonna have a lot of effects on people. Not to say AI is gonna replace everyone's job, but the effect that it has on you as a person when you see a machine do something almost as good as you that took you a very long time to craft – I don't know that everyone's ready for it. 

We as AI leaders have a responsibility to be aware of it when it happens. Because AI is not a person, and it cannot replace people for all the human things that we do. But yeah, that's been top of mind recently. I know I'm hitting you with some existential stuff, but I think we need to be prepared for it. 

Mark: No, it even goes beyond what I was asking about. I totally agree – on the use case workshops over the last two years, it's definitely gone from "oh, it's helping me write better emails" to "this could potentially save me 100 hours next month." It's cool to hear it from your perspective. You heard it from Jimmy: AI's not gonna take your jobs, but it might change the way you look at your job. 

Jimmy: I think we've hit a point where a lot of people who maybe weren't believers in AI, or maybe used AI a couple of months ago when the models weren't as good, are starting to realize, "Oh wow, this is actually really good." So I think the narrative is changing a lot as more people get exposed to this stuff. 

Mark: Jimmy, is there anything else you’d like to share, any final words of wisdom? 

Jimmy: The world is our oyster. Lots of opportunity in AI. Now it may be as simple as just letting your clients know, "Hey, I'm offering AI services," because people are starting to look. And if your clients don't know that you offer any AI services, or you're stuck in an evaluation loop for a very long time, then they might go with somebody else. So... Time is now. 

Don't Get Stuck in the Evaluation Loop 

As Jimmy said at the end of the conversation, the biggest risk right now isn't the technology—it's silence. Your clients are already looking for answers, and if they don't know you have them, they will look elsewhere. 

The time for waiting is over. Just go tell your clients you provide secure AI.