I’ve been looking back at how I’ve used AI in my writing over the last four years, and the difference is huge.

In January 2023, I published a New Year post that was built with early AI. If you want to see where we were back then, you can still find it. It reads fine, but it’s incredibly generic. Soft language and short statements but no significant risk involved. At the time I was just impressed that a model could string sentences together, which tells you a lot about where the tech was.

Fast forward to this month, and I’ve just shared a new post on the three conversations L&D teams should have in January. AI support helped shape this new post, yet it feels significantly different from the 2023 piece. There’s perspective, structure, and tension, and it reads more like an argument than an exercise in sentence creation.

What changed isn’t just the technology; it’s my standards.

Back then, the bar was: “Can AI make something readable?” Now the bar is: “Does this tool help me think better?” If it isn’t sharpening the idea, pushing clarity, or challenging the logic, I don’t use it.

That shift mirrors what’s happening in learning and development, too. Early AI use was about speed and output. Now it’s about sense-making, design and decision quality.

The 2023 post shows what AI alone could produce.
The 2026 post shows what happens when you bring AI into the thinking, not just the typing.

And that, for me, is the real story: this isn’t about replacing a human voice – it’s about improving the quality of it.