While AI has gotten better at tasks, it’s not self-sufficient, and it’s not perfect. Someone still has to prompt it, evaluate its output, edit and decide what to do with the results. That someone is you, if you choose to be.

The real story isn’t that AI is taking all white collar jobs; it’s that companies are betting that a smaller team of people who know how to use AI can outperform a larger team that doesn’t. That divide, the AI haves and the AI have-nots, is widening. And unlike most workplace advantages, this one is largely within your control.

Now is the time to integrate AI tools with your role. Learning to use AI effectively takes time and energy. In the beginning, you may spend more time on a simple task than it would have taken to just do it yourself. Think of it the way you would any new skill: slow at first, faster with practice. The big message here is to just get started.

Start with a game plan

Start with one specific task you do regularly, such as drafting a recurring report from a spreadsheet, summarizing meeting notes or researching a topic and then run it through an AI assistant. Be specific in your instructions. The quality of your results depends almost entirely on the quality of your prompt.

Good prompting means giving the AI context and clear direction. Instead of “write a summary,” try: “Summarize this client meeting in three bullet points for a follow-up email to a nontechnical audience, keeping it under 150 words.” When you find a prompt that works, save it. Keep a folder of your best prompts the way you’d keep templates. You’ll refine them over time, and they’ll become one of your most valuable professional assets.

Track what you’re doing

Keep a simple log. When you use AI, note the task, how long it took and what you did with the time saved. This helps you recognize which use cases are worth developing further, and gives you concrete evidence of productivity gains, which is the kind of thing you want to discuss in a performance review. “I automated my weekly reporting and used the hour to develop three new client proposals” is a much stronger story than “I’ve been experimenting with AI.”

Which tools to use

You may or may not get to choose your tools. My son is limited to Microsoft Copilot at work to protect company data. My daughter, a university professor, can choose whatever she wants. If your workplace has made the choice for you, that’s fine. The underlying skills transfer across platforms.

If you have flexibility, it’s worth experimenting. ChatGPT excels at creative writing and brainstorming. Claude tends to do better at analysis and longer documents, while Perplexity is really good at research. Google’s NotebookLM offers specialty items like podcasts, infographics and flash cards from text. 

One more note: Models are evolving at an astonishing pace. New features arrive in a constant stream and every week brings another announcement. Don’t get distracted chasing every update. The fundamentals stay consistent even as the models improve. Master those, and you’ll get value from whatever tools you’re using.

Find what works for you

Start with tasks that take the most time and carry the lowest stakes. A financial analyst might use AI to draft client summaries. A marketing manager might generate first drafts of campaign copy. A project manager might turn rough notes into structured action plans. Once you’ve proven value in one area, move to the next.

The professionals most protected from the wave Dorsey is predicting aren’t the ones who know every AI tool. They’re the ones who have figured out how AI fits into their specific work and have the results to prove it. Right now, the bar is still low enough that starting today puts you well ahead.

Leslie Meredith has been writing about technology for more than a decade. As a mom of four, value, usefulness and online safety take priority. Have a question? Email Leslie at asklesliemeredith@gmail.com.