There’s no shortage of conversation right now about what AI will look like in the years ahead. Whether it will reshape entire industries or replace roles we thought were safe. What our workflows will even look like in five years. Elon Musk recently suggested we might not even need to save for retirement because AI will create a world of abundance. Underneath all of that, a quieter concern about who’s accountable when these systems fail.These conversations matter. But as we head into a new year, I keep coming back to something more immediate: before we speculate about where AI is going, it might be worth pausing to examine where we actually are with it.

Recently, someone in one of our courses said to me, “Audry, I feel like I’m using AI for everything now.” That comment stuck with me. It’s not wrong. In many ways, it’s a sign of how quickly AI has become embedded in our practice. When many of us first started integrating AI, it felt like we were reclaiming time. But after that initial phase, a different question starts to surface. Not can we use AI for this, but should we? And more importantly: is it actually making our work better, or just making it busier?

Using AI frequently is not the same as using AI intentionally. For a lot of people, AI has become become another layer in the workflow, one that involves constant prompting, re-prompting, clarifying, copying, pasting, and refining. If that sounds familiar, it’s worth asking honestly: where is AI genuinely saving you time, and where might it actually be adding friction? Where is it helping you think, and where might it be replacing the thinking you actually want to do yourself?

This is why I’ve been encouraging people to move beyond one-off prompts. Bots and custom assistants let you embed your context and expectations once, rather than re-explaining them every time. That shift alone can reduce the cognitive load significantly. AI starts to feel less like something you’re managing and more like something that’s actually supporting you. For those with access to more advanced tools, agents take this even further. Instead of guiding every individual step, agents can handle multi-stage processes with less manual intervention. Done well, this lets you design workflows that match how you actually think, so you can spend less energy on the tool and more on the work that actually requires you.

We’re also seeing new possibilities inside platforms we already use. Google’s Opal, for example, lets you build simple apps using natural language. No code, no technical background required. A year ago, if you wanted a tool that did something specific for your workflow, you either found one that came close or you didn’t have one. Now you can just describe what you need and build it yourself. The broader point is that we have more agency than ever to shape our tools around our needs, rather than bending our work around the tools.

The future of AI will keep evolving whether we’re ready or not. But how we choose to work with it today is something we can shape. Heading into this year, the real opportunity might not be in doing more with AI. It might be in doing less, more intentionally.