,

My Thoughts on AI – Part 5 – When All You Have Is a Hammer: Choosing Your AI

You know the old saying: when all you have is a hammer, everything looks like a nail. For a while, that was me with AI. 

Welcome back to our series on AI adoption for facilities leaders. Today I want to tackle a question I get all the time: which AI should I be using? 

Fair warning: this post will be more timely than timeless. In the AI world, today’s best tool for a task can be tomorrow’s has-been. 

Start with what you’re allowed to use. 

If your organization says you use one AI, that’s the AI you use. I don’t recommend working outside the boundaries your IT department sets. I hear from a lot of clients that they’re limited to Copilot in a Microsoft shop or Gemini in a Google shop. That’s okay. These tools are improving quickly, and some, like Copilot, now let you tap into other vendors’ models behind the scenes. 

Stay close to the people who make the rules. 

AI policies are evolving almost as fast as the tools. Get to know your IT decision makers so you know what’s coming down the pike, and so they know there’s demand. Remember: no today often means not yet. 

If you do have a choice, it’s harder than it used to be. 

When ChatGPT burst onto the scene, it was the clear leader. It’s where we started at Roth IAMS, once we sorted out IP protection. Then Claude made a big leap forward in coding with Opus 4.5. As we built more internal tools and apps, we moved almost entirely to Claude. 

I held out personally for a while. I had ChatGPT trained and I didn’t want to start over. Then I discovered I didn’t have to. I built a single document that taught Claude everything ChatGPT had learned from months of working with me. 

Build a toolbox, not a favorite. 

We’re still mostly a Claude shop. But lately we’ve been reaching for other tools when they’re best at a specific job. For example, we use Perplexity to research organizations as part of our prospecting. Different jobs need different tools, just like in the physical world. 

Find your AI champion. 

Someone needs to keep an ear to the ground on how each model is evolving. If that’s you, make it part of your job, not a side hobby. If it isn’t, get to know whoever it is. They’ll save you a lot of trial and error. 

Your homework: 

Pick one task you do every week, like summarizing meeting notes, drafting a board memo, or reviewing a vendor proposal. Confirm which tools you’re allowed to use. Then run the same task through two of them and note which did better, and why. Congratulations: you’ve just started your toolbox. 


Next post I’ll take you through my personal journey of building an AI content collaborator, and why I didn’t have to start from scratch.   

   

Published on

22 October 2026

Under

,

Further Insights

At Roth IAMS, we take great pleasure in sharing our stories and knowledge