Four of the five things a great coach does, you can get out of your phone tonight. The fifth is the reason I still pay humans.
At 17, fresh out of rehab, somebody handed me a piece of paper and told me to write a transition plan for my new life.
One of the questions asked what I actually wanted to achieve. I had never written a goal down before. I had been taken from home at 11, sent to boarding school, and spent most of my teens inside the system, so up until that afternoon my whole plan was just don’t fall apart.
I wrote something insane for a kid in my position. Earn $100,000 a year. Then I wrote: buy a house.
At 21 I made my first $100,000. That same year I bought my first house.
If I had written “stay sober,” none of it happens. The goal wasn’t big enough to pull anything out of me. Somebody had to sit across from me and push me past what I could see, and I have spent hundreds of thousands of dollars on coaches ever since because that push is the whole game.
Here’s what changed: four of the five things those coaches do for me, AI will now do for free. Below are all five, in the order you have to run them, with the exact prompt for each.
1. Dream first, then put a number and a date on it
A coach doesn’t make you slightly better at what you’re already doing. They prepare you for your potential, which means the first thing they do is make you set a bigger goal than you’d set alone.
So start with the dream. Let yourself think a full decade ahead and name five to ten outcomes across work, money, health, growth and relationships. Don’t ask AI to write these for you. This part is yours. AI’s job is to catch you being cautious.
Then convert each dream into a goal, and a goal needs exactly two things: a number and a date. “I want to be in shape” is a dream, and dreams are comfortable because you can’t fail one. “Down 20 pounds by December 31” is a goal. Once every dream has a number and a date on it, you can back it into 10 years, 3 years, and this year, and the plan builds itself.
Open a dictation app and read this out loud to your AI:
You are my strict mentor and your job is to make me dream bigger. Ask me how I want my life to look in 10 years across work, money, health, relationships and personal growth. When an answer sounds conservative or too easy, call it out and push me. I want to 10x my life, not 2x it. Then turn every answer into a real goal with a number and a date. If I say something vague, don’t let it slide. Finally, work backwards into what that means in 3 years and what it means this year.
It will argue with you. Every time I’ve run it I get “that’s not specific enough” and “I don’t buy that.” That argument is the point.
2. Grade yourself before AI tells you you’re doing great
Now you need the truth about where you actually are.
Every Friday I open my calendar and do my non-negotiables. One of them is a table with seven columns: health, relationships, love, work, money, fun, personal growth. I score myself 1 to 10 in each, highlight the two lowest, and commit to moving those two the following week. It takes 10 minutes, and it is the single biggest reason nothing in my life turns into a crisis. I always know where I stand.
Ask most people how their health is and you get “pretty good.” Based on what? There’s a name for that gap. Dunning-Kruger: almost everybody rates themselves above average, which by definition can’t be true. Ask me how good a driver I am and I’ll tell you I’m the best. Obviously not. We are worst at this wherever there’s no system keeping score.
The fix is to score behavior, not feelings. What does a person who hits your 10-year goal do every single week? Someone chasing an athletic body trains hard and stays in a deficit. Both of those are scoreable. Skipped the gym, that’s three or four points gone. Ate badly, another three. Once you’re handing yourself twos and threes, the fantasy dies and you can see it.
This is exactly where AI is dangerous, because left alone it’s a sycophant. You’re a genius, you’re so close, keep going. In 2025 OpenAI pulled a ChatGPT update specifically because it was too agreeable, nodding along instead of telling people the truth. So don’t ask it how you’re doing. Set the standard first, then make it grade you against the standard.
You are my ruthless coach, not my assistant. Tell me the truth instead of trying to please me. Use my 10-year goals to define what a 10 out of 10 looks like each week in every area we talked about. Turn that into a simple grading system built on the non-negotiables that actually move me toward those goals. Every Friday, interview me about my week, score me 1 to 10 in each area, and tell me my two lowest scores.
I did that by hand for years. I handed it over because I shouldn’t be the one grading me.
3. Your brain will hand you an easier problem to solve
You’ve got two low scores. This is the moment it usually falls apart.
A woman at one of my events, health industry, deep into AI and automation, told me she was building a state-of-the-art customer support system because support was her bottleneck. The easy move would have been to help her build it. Instead I asked three questions. Where are you now? Where do you want to be? What do you think the real problem is?
Ten minutes later the real answer showed up. Support wasn’t the problem. She didn’t have enough practitioners to serve the clients she already had. She was about to spend months building a beautiful solution to a problem she didn’t have.
Psychologists call this attribute substitution. Your brain quietly swaps the hard question for an easier one that feels just as productive, so instead of “why do I keep skipping the gym” you get “what AI tool should I build for my business.” You spend a week on it. Nothing has changed. That’s how smart people work for a year and land in the same spot.
And AI makes this worse, not because it’s bad but because it’s good. Give it the wrong problem and it will research it, plan it and build it faster than you ever could. Congratulations, you’re now wrong at scale.
So don’t let it solve anything until it has argued with you about what the problem actually is.
My two lowest scores this week are X and Y. Don’t let me fix anything yet. Interview me about how I actually handled those two areas this week. Go several rounds, and each round dig into what you think the real reason is behind my last answer, especially the part I can’t see. Keep going until you hit something I haven’t mentioned. Then give me one change for each that would move the score the most, and explain why you picked it over the obvious answer.
Say your low score is your marriage and your instinct is “plan more date nights.” Run those rounds and you might land somewhere else entirely, like the fact that you don’t clean up after yourself and she’s been asking for a year. How do I know? Same conclusion, and it took me way too long to get there. Or you think you need more gym time and it turns out you’re not sleeping enough to recover from the training you’re already doing.
4. A number with no stakes attached is just a number
Here’s the fourth job: a great coach makes you feel like you’re winning before you’ve won, by measuring progress and attaching something real to it.
A number in a spreadsheet means nothing on its own. It needs an emotion bolted to it, either the reward you want or the pain you’re avoiding. I call that reward and stakes, and the stakes matter more, because you will do more to avoid a loss than to chase a win.
Jen on our team set a fitness goal, a weight she hadn’t seen since school. If she’d just promised herself a trip to Bali, she wouldn’t have hit it. Instead she set her own stake: miss the goal, lose her spot on the team. On the days she didn’t feel like training or tracking, she did it anyway. She stepped on the scale every morning. The trip would never have done that.
Pro tip: I tie money goals to purchases and experiences, and health goals to family experiences. If the family vacation is riding on me getting lean, my family stops pushing dessert at me and starts backing me at the dinner table, because it’s their trip too.
Then run it at three levels. The year is the big one, with a real reward and a real consequence. Last year we set out to take our subscriber base from 3.2 million to 8.9 million, and we put our cars on the line: my GT3RS, Todd’s Taycan, Sam’s GT4. Quarterly milestones under that. Daily activity per person under those. It came down to December 30th with a genuine chance we’d miss, and Betty, who runs our house, was ready to list all three cars. No negotiation. We got there, and I got to surprise Todd and Sam with watches instead.
I know it’s intense. I’m also not going to teach you this and not live it.
Take the annual goals we set and break them into one milestone per quarter and the daily actions that get me there. Every level needs a specific number and a date. Then interview me to pick a real reward and a real consequence for the year, plus a smaller reward for each quarter. Push me on the consequence: if it’s something I could ignore, tell me and make me choose again. Then build me a spreadsheet dashboard I can open every day showing yearly progress, quarterly progress, today’s non-negotiables, and the number of days left.
5. The one thing AI can’t do
That’s four jobs done. Dream bigger, tell me where I am, find the real problem, keep score. There’s a fifth, and it’s the reason anyone serious still hires a human.
This channel exists because of a conversation with my coach. They looked at what I’d shared and said: you need to build a media team, you need to become a professional at this, and it needs to happen before our next session or I stop coaching you. I would never have felt that weight from a spreadsheet. Somebody I paid, whose opinion I valued, who had already done it, told me to do the thing and told me what it would cost me if I didn’t.
AI will never be able to do that. You can let it down at 11pm and nothing happens to you.
Knowing what to do was never the bottleneck. There will be a month where you don’t want to do any of it, and what carries you through is a real commitment to a real person who will come looking for you. The best version isn’t one person either, it’s a group. I have coaches for different parts of my life: one who teaches me to live more openly, one who helps me with my kids, one for my health. All of them have already been where I’m trying to go. That’s why I put my coaching members into groups, so they’ve got me pointing them at the right work and each other for everything else.
So go find one. Someone who has actually reached the goal you’re chasing, not a friend, not family, not somebody one step ahead of you. Tell them the goal. Meet on a schedule instead of whenever you feel like it. And when you’re in the room, ask the question they’re actually there for: in my situation, what would you advise me to consider?
I use AI to prep for those meetings so we don’t spend the first 20 minutes on background. I show up with the plan and we work on the parts that need a human.
Based on everything we’ve built – my goals, my scorecard and the real problems we found – give me a list of coaches, programs and communities that could actually help with this specific problem. Ask me anything you need first. Then give me the questions to ask before I pay for anything, plus the red flags that tell me they’ve never actually done it themselves.
Run it again
Do all five and you don’t just get an answer, you get a machine you can run again. Hit the goal, start over, point it at the next area of your life.
Jordan didn’t need Phil Jackson to explain basketball. He needed someone who would say the thing nobody else would say. Most people never go get that person, and then they call it being stuck.
Go get the person.
-DM