If you're finding yourself spending more time reviewing and updating poor-quality work, you're not alone.

It’s a sign that you (and your team) need a better operating system.

That's what this week's AI at Work is all about.

You'll get the practical operating model and shared habits (The Driver Model and The AI Habit Stack) you need to set expectations, protect human judgment, and raise output quality without banning AI.

Plus, we'll dive into the one thing everyone wants to know, “how do I get AI to write like me.”

Ready to level up your work?

Read on.

When AI Starts Making Your Team Sloppy

Last week after a keynote a manager pulled me aside complaining that her team is relying too heavily on AI.

“It’s making them lazy,” she lamented.

She wants people using their judgment, building their skills and standing behind their work and she’s close to restricting AI use to protect those standards.

I understand the instinct.

When quality starts slipping, the fastest reaction is to remove what's causing the issue, but if you ban AI, people will use it anyway—they just won’t admit it.

This is the same problem I work on with teams at Section.

I teach a framework to leaders and operators inside Fortune 500 organizations who are actively rolling out AI. It gives them a concrete way to set standards and coach usage, while employees develop a more deliberate approach to how they apply AI to their work.

I call it the Driver Model.

Here’s how you can put into practice:

1. Intent First
​Define the outcome you are driving before you engage the tool. Whether the goal is clarity, decision quality, speed, learning or scale, a clear goal shapes how AI contributes to the task and prevents you from spending hours on aimless prompting.

2. Human Judgment Anchors
​Establish where your judgment remains primary across the entire workflow. Strategy, tone, tradeoffs, risk assessment and final decisions are yours, with AI supporting analysis and exploration along the way.

3. AI Leverage Zones
​Identify where AI consistently creates meaningful lift. Drafting, synthesis, analysis, pattern discovery, and administration often qualify and free capacity for higher-value thinking, which is where your experience earns its keep.

4. Feedback Loops
​Review output quality on a regular cadence to maintain standards. Track what worked well, what reduced quality and what should improve next to reinforce learning and consistency over time.

For managers, this becomes a practical coaching mechanism that shapes expectations, guides review and anchors feedback. It also gives you a neutral way to talk about standards without turning every review into a debate.

If you’re an individual contributor, you can use this as a self-regulation tool that helps you maintain your standards. If you ever catch yourself thinking, “This is good enough, I’ll just send it,” that’s usually the sign to slow down.

If you notice your colleagues falling into the AI slop trap, share this framework with your manager and suggest rolling it out as a team standard. A shared operating model raises output quality, reduces rework and sets clear expectations.

Everyone will be happier, especially the person who no longer has to rewrite half the drafts before they go out the door.

Build Strong Work Habits with AI

Inside most teams, working habits shape how projects progress, problems get solved, and how work moves through the day. AI amplifies those habits, which means the patterns people repeat every week determine whether leverage compounds (or friction creeps in).

If you want the Driver Model to stick, it has to live inside daily work—the briefs, drafts, follow-ups and decisions we all make.

I teach teams to think about this as a simple AI Habit Stack, a handful of behaviors that help you stay on track.

Here it is:

Input Discipline
​Good output starts with good input: clear context, constraints, and examples shape the quality of your output and reduce “cleanup”. If you save 15 minutes on a task you do five times a week that adds up over time.

Thinking Discipline
​Strong users treat AI as a thinking partner, but still own the conclusions. They pressure-test assumptions, layer in their experience and assess multiple angles before moving forward.

If you ever catch yourself saying, “Well, Copilot said…,” that’s a sign your thinking has left the building. This leads to bad decisions and the likelihood you’ll be pursuing something the thousands of other people who also received the same advice that day.

My advice? Ban that phrase from your vocabulary!

Execution Discipline
​Reusable prompts, documented workflows and consistent naming conventions reduce mental overhead and make work easier to share, review and improve. Take the time to build a shared repository of workflows and prompts. You can use an Excel file, the Copilot prompt library or a Notion database. Just make sure everyone has access.

Operating Discipline
​Habits stick when you run them as part of an operating model. Without one, even strong thinking and execution can drift into reactive work.

That’s why I created a dedicated ChatGPT Project I call Peak Performance. It delivers a daily rundown every morning and acts as my planning and execution layer.

It helps me:

• Review upcoming meetings and surface likely preparation needs
• Flag action items coming through email so nothing slips
• Structure deep work blocks intentionally
• Manage energy across the day instead of burning it early
• Close the loop with lightweight reflection and adjustment

If you want to see how this works in practice, you can download my Peak Performance Blueprint and install it inside your own AI tool.

If you want a lightweight way to reinforce these habits, here are a few weekly reflections I use with teams (and myself):

Weekly AI Habits Check

• Where did AI improve clarity, speed or decision quality this week?
• Where did output miss the bar I expect?
• What task deserves a stronger AI workflow?
• What judgment should stay fully human-owned?

Five minutes of reflection keeps habits tight and prevents drift. It also gives managers a way to coach usage without hovering, which everyone appreciates once the novelty wears off.

How to Build Your AI Workbench

Imagine a week where every project wraps on time, thoughtful emails get written in minutes, and AI absorbs the admin work you prefer to avoid. Most people already have the tools to make this happen, but don’t have them connected to their systems or workflows.

As an example, I’ve been refining my own setup over the past few weeks, and two changes have had an outsized impact on my day.

The first is how I transfer ideas from my head and onto the page.

I use Wispr Flow to dictate my working drafts because I can speak at 140 words per minute, versus typing at 90. I also make fewer errors when I talk, which reduces cleanup and keeps my momentum going.

The second change is where I use ChatGPT.

I spend a lot of time inside my browser working across tools, dashboards and platforms. I now use ChatGPT’s Atlas Browser for things like research, troubleshooting and filling in forms. It will even look up shopping discount codes and test them before I make the purchase!

One of my favorite examples came from building an automation in Make.

As a non-coder I found the UX dense and unintuitive. When I tried building the automation unassisted, my frustration levels redlined. I would have quit the entire process but I moved the window into Atlas and asked for a step-by-step guide.

The back-and-forth felt closer to having a patient teammate sitting next to me and pointing out where I had gone wrong and how to fix it rather than searching documentation or guessing my way forward.

These are just a few ways I’m optimizing my AI workbench but hopefully, they inspire how you refine your own system.

Here’s a quick checklist to get you started:

1. Speed of Capture
​How quickly can you get ideas, questions and drafts out of your head and into a workable form? Voice, shortcuts and lightweight capture tools often create the biggest gains because they remove small delays that add up across the week.

2. Context Awareness
​How well does your AI understand what you’re looking at, working on and deciding right now? Tools that live inside your workflow (like Atlas, Copilot and Gemini) reduce the need to translate problems into long explanations and screenshots.

3. System Connections and Personalization
​Are your AI tools connected to the systems you already live in, such as email, calendar and messaging? Do your personalization settings reflect how you work? Custom instructions, memory settings and default preferences reduce repetitive prompting and improve outputs.

This is also where it pays to review privacy and training data controls so your usage aligns with your comfort level and company policy. As an example, I always make sure I have the “training data” toggle turned off.

4. Flow and Focus
​Does your setup support sustained focus across long work blocks, or does it force you into constant context switching? Fewer hops between tools and clearer defaults preserve your energy and reduce fatigue.

A Quick Self-Check

If you want a quick way to assess your setup, ask yourself these questions:

• How fast can I capture a rough idea or create a draft?
• How much explaining does my AI need to understand the problem?
• Are my core tools connected and personalized?
• Where does friction or context switching slow me down?
• Where does my attention get fragmented during deep work?

You don’t always need a full rebuild. Even if you focus on one or two of my checklist items you create an immediate lift.

Workflow of the Week

How to Get AI To Write Like You

I've been having a lot of conversations lately about AI detectors and whether or not they actually work.

Short answer? They don't.

A while back, the internet was freaking out because AI detectors claimed The Declaration of Independence was 99% written by AI. When I tested it this week, 50% of them still agreed.

While a lot of people rely on these AI detectors to check if people are cutting and pasting from ChatGPT, these detectors are notoriously bad at detecting whether or not something was written by AI.

The reason is that AI detectors look for patterns like organized structure, formal wording, and flow. The more "polished" your writing is, the more likely it gets flagged.

Meanwhile, humans can spot AI writing from a mile off. Em-dashes and phrases like "game changer" and "dive into”, generic structure that could've been written by anyone. And my personal pet peeve, “it’s not this, it’s that”.

You might be trying to fix this with your prompts, and yes, if you tell Gemini to avoid these words and patterns, you'll get writing that sounds less like AI, but it still won’t sound like you (or your brand).

If you want AI that writes the way you do, here’s a simple way to set it up:

1. Analyze your best writing to identify your style and tone. Identify sentence patterns, signature phrases, and word usage.
2. AI loves examples, so feed it 4-5 pieces of your writing that sound most like you.
3. Create a Custom GPT, Project or Gem and give it explicit instructions, including your tone and style, and what NOT to do. As an example, I have a list of 117 banned phrases and words and a great hack for eradicating those em dashes.

Unfortunately, there's no shortcut to make AI write exactly the way you do.

Even after you build out this framework, you'll still need to spend time reviewing and refining. You need to help the AI understand your context, how to prioritize your points and you can’t outsource the editing step. No matter how well you customize your AI, it still needs that final check.

That said, if you take the time to set it up correctly, you'll save yourself hours and you won't have to worry about AI detectors.

Want a full walk through? Head over to my YouTube channel here.

Want to Level Up Your AI Game?

If your team is ready for a hands-on AI strategy session, my custom-designed workshops are built to uncover the workflows that can save you hours every week.

Prefer to start small? My YouTube channel is packed with quick, practical “how-to” videos that show you exactly how I use AI tools for marketing, content, and automation.

Planning an event or conference? I deliver high-energy AI sessions that engage audiences and leave them with actionable strategies they’ll talk about long after the event. Book me for your event here.​