7 Ways to Be Good at AI People will think you’re an AI

Sam Liang I was horrified when I admitted my method for recording the interview: launching the voice memos app on my iPhone and manually transferring the transcript to a Google Doc. The CEO of Otter, a transcription service for analyzing meetings, looks at me as if I’m trying to connect to a video chat with a rotary phone. He thinks, of course, that I should turn into an Otter. Maybe he’s right.

It’s all part of a new identity at work (and perhaps at home): native AI. Time-saving productivity tools like next-generation note-taking, to-do agents, and chatty inbox assistants are increasingly popular, invading every nook and cranny of our digital lives. While it’s important to keep concerns about safety and hallucinations top of mind when using any AI feature, early adopters develop a fluency that will likely pay dividends for years to come.

Being an AI native — or a “doer,” as AI natives say — means remaining adaptable to new experiences. Aside from transcription failures, I’ve embraced experimentation, from creating AI-powered podcasts to letting Claude organize my desktop files. (I talked about some of this in my newsletter series last year, AI Unlocked.) If you want to get so good at using AI tools that your coworkers start wondering if blood or tape cables are running under your skin, here are my seven tips for the rise of AI.

1. Kill your chatbots

ChatGPT It’s 2022. These days, all the cool kids care about is Codex. Your eyes may, rightfully so, glaze over at the mention of “AI agents,” but compared to anything on the market even a year ago, software automation tools like Codex and Anthropic’s Cowork are much better at taking control of your computer and completing tasks. Don’t waste your time messing around with one chatbot when you can command a whole army of them.

2. Switch to voice mode

Oh, are you still writing out everything you want your AI tools to do, Boomer style? That’s nice. But trust Otter’s Liang: “The sound will become more dominant going forward,” he told me. “People hate writing.” (He warns against this Ia journalist, probably doesn’t hate writing, and that’s mostly true.) This step is primarily for inputs, not necessarily outputs. I rarely use ChatGPT’s voice-only mode, for example, but often speak a prompt into my phone and then browse through the written output.

3. Build a sandbox

Even though the agents were really good now, the rascally little demons were still capable of destroying everything without proper limits. (Earlier this year, a Cloud-backed agent deleted a startup’s entire production database and backups.) So if you’re ready to let an outside entity control your computer, you need to spend an afternoon researching everything these tools can do and set up some custom folders with the files you want them to access.

4. Give it everything you have

With apologies to privacy-conscious security writers, it’s simply that the more data you share with AI, the more personalized the output becomes. Jo Barrow is chief of staff at Granola, one of Otter’s competitors, and she explains it this way: “I have a personal operating system, which is a series of files on my computer that my AI lives in. Whenever I ask questions, all that context is there, and the agent can go and figure it out. I don’t need to repeat myself over and over again.” Fair warning: It’s still best to have sensitive conversations without a permanent record.

5. Create an impersonator

Barrow told me that she stores all of her Slack messages in a document to let bots know what she looks like on that platform, and does the same for her email inbox and social media accounts. “People are using AI to improve the tone of their voice,” she says. “There’s only so many times you can say, ‘Okay, a little bit warmer.’ Well, a little less formal.” This is a huge time sink. Creating these clues for agents to follow won’t exactly replicate your voice, but they can prompt the bot to output something at least closer to your cadence and tone.

6. Think across teams

Data is powerful, and adding more of it from the people around you can further improve AI tools. Consider your coworkers: “Many people use a meeting note taking tool now, but they still use it at the individual meeting level,” says Liang. He touts the “knowledge engine” that Otter can create when… The whole workplace Buys, from the engineering team to the marketing department. You can also do this at home: if family members pour different notes from their day into one shared AI tool, it will provide more insights than isolated use.

7. Learn how to escape from prison

Successfully using AI tools in 2026 doesn’t require writing, I mean to talk—Perfect claims. However, starting more complex tasks with a creative and well-calibrated request can be challenging. Experiment with the wording, especially if you run into unexpected guardrails that prevent output. I recently tried to get a bot to send email addresses to a variety of subject matter experts, but it refused to deliver. But when I started a new conversation and shared details about why I wanted this information (for reporting purposes, not stalking, of course), the list branched out.

What does she tell you?
Tell us what you think about this article in the comments below. Alternatively, you can send a letter to the editor at [email protected].