7 Ways to Master AI So Well People Think You’re AI

Sam Liang reacts with visible dismay when I reveal my interview recording process: using the iPhone’s Voice Memos app, then manually copying the transcript into a Google Doc. The CEO of Otter, a meeting transcription platform, looks at me like I tried to connect to our video call with a rotary phone. He insists I should switch to Otter. He’s probably correct.

This shift is part of a broader identity emerging in the workplace and possibly at home: the AI native. Next-generation note-taking tools, task-oriented agents, and chatty inbox assistants are exploding in popularity, embedding themselves in nearly every corner of our digital routines. While security risks and hallucinations remain critical concerns with any AI feature, early adopters are developing a fluency that will likely pay off for years.

Being AI native, or “agentic” as some call it, means staying open to new experiences. Despite my transcription missteps, I’ve embraced experimentation, from generating AI podcasts to letting Claude organize my desktop files. If you want to become so skilled with AI tools that your coworkers wonder whether blood or ribbon cables runs through your veins, here are seven strategies for AI-powered ascendance.

1. Kill Your Chatbots

ChatGPT feels positively ancient. The cool crowd now gravitates toward Codex. You might roll your eyes at “AI agents,” but tools like Codex and Anthropic’s Cowork are vastly better than anything available a year ago at actually taking over your computer and completing tasks. Stop wasting time with a single chatbot when you can command an entire army of them.

2. Go Voice Mode

Still typing every prompt for your AI tools, Boomer-style? That’s quaint. Liang from Otter insists, “Voice will become more dominant moving forward. People hate writing.” (He acknowledges that journalists like me probably don’t hate writing, which is mostly true.) This shift focuses on input, not output. I rarely use ChatGPT’s voice-only mode, but I often speak a prompt into my phone and then skim the written result.

3. Build a Sandbox

Even though agents are genuinely capable now, these mischievous tools can still wreak havoc without proper boundaries. Earlier this year, a Claude-powered agent deleted a startup’s entire production database and backups. If you’re ready to let an external entity control your computer, spend an afternoon researching what these tools can do and set up dedicated folders with the files you want them to access.

4. Give It Everything You’ve Got

With apologies to privacy-minded security writers, the more data you share with AI, the more personalized the outputs become. Jo Barrow, chief of staff at Granola, an Otter competitor, explains: “I have a personal OS system, a series of files on my computer where my AI lives. Whenever I ask questions, all that context is right there, and the agent can figure it out. I don’t need to repeat myself over and over.” A fair warning: sensitive conversations are still best kept off the record.

5. Create an Impersonator

Barrow says she dumps all her Slack messages into a document so bots know how she sounds on that platform, and she does the same for email and social media. “People use AI for finessing their tone of voice,” she says. “There’s only so many times you can say, ‘OK, a little bit warmer. OK, a little less formal.’ That’s a big time sink.” These guides won’t fully replicate your voice, but they can nudge the bot toward something closer to your cadence and tone.

6. Think Across Teams

Data is powerful, and adding more from people around you can further enhance AI tools. Consider your coworkers: “Many people are using a meeting note-taker now, but they’re still using it at the individual meeting level,” Liang says. He promotes the “knowledge engine” Otter can create when an entire workplace buys in, from engineering to marketing. You can even do this at home: if family members pour notes from their day into a shared AI tool, that yields more insights than siloed usage.

7. Learn to Jailbreak

Successfully using AI tools in 2026 doesn’t require writing perfect prompts. Still, starting complex tasks with a creative, well-calibrated request can be crucial. Experiment with wording, especially if you hit unexpected guardrails. Recently, I tried to get a bot to send me email addresses for niche experts, and it refused. But when I started a new chat and explained why I wanted the information for reporting purposes, it handed over a list.