5 Things Solo Founders Do With AI That Every Leader Should Steal

Your team isn't your AI advantage. You are.

A quick note before we start.

This is one half of a paired conversation with Joel Salinas, who writes Leadership in Change for executives putting AI to work and works as a fractional AI Officer inside companies actually running these rollouts.

Joel and I traded notes to create this collab.

He sees it from the org side, watching real AI pilots succeed or stall.

I see it from the solo founder side, where there is no leadership team, just you and the tools.

Working solo forces a relationship with AI that working at scale tends to suppress. Here are the five places that gap shows up and how to close it.

I run a portfolio of small businesses solo. A newsletter (No-Code Exits) with around 13,000 readers. A done-for-you AI agency. A few products. A few partnerships.

No permanent team. No assistant (unless AI counts). No one to answer to but myself.

The only reason any of this works is how I use AI, and the way I use AI looks almost nothing like how leaders at real companies use AI.

When you're solo, AI isn't a productivity tool you bolt onto an existing operation. It is the operation. There's no team to delegate to, no procurement chain to escalate through, no committee to socialize a decision with.

You either figure out how to use AI as a thinking partner, an engineering team, a peer, and a critic, or you fall behind.

Leaders inherit the opposite problem. They have all of that infrastructure, which sounds like an advantage. The peer reviewer agrees with you because you sign their performance review. The engineering team is busy. The strategy session generates documents instead of tests.

Your environment is plagued by busy work and bureaucracy.

The five habits below are the five places solo founders have built a working relationship with AI that most leaders haven't, because most leaders have never had to. Not because solo founders are smarter. Because their environment forced something the leader's environment is actively suppressing.

The Leadership Trap (Read This First)

Most leaders adopting AI in 2026 are doing it through their organization.

Pilot programs. Vendor selection. Enablement training. A Slack channel where someone pastes ChatGPT outputs.

That's not adoption. That's procurement with extra steps.

The leaders who actually compound advantage from AI are the ones who build a personal practice with AI before they roll out an organizational one. Not because the org program is wrong, but because if you don't have your own working relationship with the tool, you can't tell which parts of the org program are working and which parts are broken.

You become dependent on your team to tell you whether AI is working in the org. And your team is incentivized to tell you it's amazing.

Solo founders had to figure out the personal practice first because they had no choice. Leaders have a choice. Most pick the generic rollout because it's the one their team is asking them about. They skip the only step that would have made them dangerous.

Here are the five habits that make you a real leader using AI and not an "AI Leader."

1. They Use AI as a Sparring Partner

The biggest information problem in solo work is that no one disagrees with you on purpose. There's no co-founder to say "this is wrong." No team meeting where someone with skin in the game pushes back. Just you and your own bias.

And if you let it, AI will agree with everything you say.

The solo founder fix is to make AI a sparring partner, on purpose. They set it up explicitly:

- Argue against this strategy.
- Find the strongest version of the case I'm missing.
- Tell me what a senior operator at a competitor would point out that I'm too close to see.

Done well, this is the closest thing a solo founder has to a board meeting.

Leaders have the opposite information problem and somehow end up in the same place. Your environment is full of people who could disagree with you.

Most won't.

The CFO who reports to you will not tell you your forecast is fantasy in front of your COO. The VP angling for your job will not point out the strategy hole until the QBR after you've committed. Your direct report doesn't know how to disagree with you in a way that doesn't cost them. They've calibrated their honesty to your tolerance. You've calibrated your tolerance to your comfort. The system has reached a quiet equilibrium where everyone agrees with the leader, and the leader thinks the strategy is sound.

AI is the one peer in your environment with no incentive to manage you. Use it the way the solo founder does.

Tell it to disagree. Give it the brief your actual board would receive, and ask it to play the toughest member. Ask what your CFO would say if she didn't report to you. Ask the question your chief of staff is too polite to ask.

The most expensive thing on a senior leader's calendar is the meeting where everyone agreed.

The first time I ran this exercise on my own work, the model surfaced three assumptions I had never said out loud. Two of them were wrong.

I hadn't said them out loud because I had no peer to say them to. I'd been carrying them as background until the model surfaced them and asked me to defend them. That's the move. Solo founders run that exercise weekly because they have to. Leaders should run it weekly because their environment won't.

2. They Build the Small Tool Instead of Staffing the Project

When a solo founder hits a recurring annoyance with the same bit of work that eats two hours every week, the same data reformatting, the same five-step research workflow, the default move is to build a small tool that handles it forever. A few hours of building eliminates the next year of friction. The math is obvious.

Leaders almost never do this. They do something else: they staff the project.

They put it on a roadmap. They scope it. They tell someone in ops to "look into it." They brief a vendor. The project enters the institutional pipeline, and somewhere between scoping and shipping it picks up six months of latency, three layers of stakeholder review, and a price tag that justifies none of it.

The lesson isn't that leaders should stop staffing projects. Some projects need to be staffed. The lesson is that leaders are operating with two pieces friction. The org-scale problem worth a team, and the personal-scale problem worth a Saturday.

If you do the same prep work for every QBR, that's a personal-scale problem. If you rewrite the same ten paragraphs in every reorg memo, that's a personal-scale problem. If your inbox is an unscalable nightmare every time you come back from PTO, that's a personal-scale problem. None of these belong on a team's roadmap. All of them are worth four hours with AI to build the small thing that handles the next twenty instances.

A few months ago I built a tool that takes my newsletter analytics and writes my next-issue brief in fifteen minutes. It's unglamorous. It's not a real product (yet). It's already saved me countless hours this year.

The compound version of doing this enough times isn't the tools. It's the orientation. Once you've built three small tools yourself, you stop seeing personal friction as the cost of being busy and start seeing it as leverage you haven't claimed yet.

The personal-scale problem is the leader's blind spot. It doesn't deserve a team. It deserves your Saturday.

That's the shift solo founders are forced to make. It's available to leaders by choice. Most don't take it.

3. They Externalize Their Judgment So It Travels Without Them

Solo founders eventually figure out that the limit on a one-person business isn't effort. It's that their judgment lives in their head, and a head can only be in one room at a time. The unlock is putting the judgment somewhere else.

The first time you do it, it's boring and tedious.

You write down the criteria you actually use to decide things. You record your taste, your style, your tone. You codify the principles you apply when you evaluate a piece of writing, a partnership offer, a customer complaint, a hire.

Then you feed all of that into a context document, a series of prompts, or a small custom agent. When the next decision shows up, you ask the system "what would I do here, given what I've told you?", and it hands you back a draft of your own thinking that you can sharpen. Faster than thinking from scratch. Often sharper, because the system doesn't get tired and doesn't have a meeting in ten minutes.

Leaders should steal this immediately, and most haven't. The reason is interesting: leaders assume the externalization tool already exists in their org.

The brand book holds the voice. The strategy doc holds the priorities. The values poster holds the decision criteria. Surely the team can run on those without you.

They can't. Brand books and strategy docs are written by committees for legibility. They are not your judgment.

Your actual judgment includes the dozen unwritten rules you apply when you read a draft, the seven things that automatically disqualify a candidate, the four moves you'd never approve and the two you'd green-light without asking.

The org artifacts don't hold any of that. Your team is reverse-engineering your judgment from the meetings they happen to attend, which is why every decision ends up on your desk, and why you're the bottleneck.

The solo founder version of this is "externalize so I can scale." The leader version is "externalize so my team can run." Same habit. Bigger payoff at scale.

If your team has to ask you what you'd do, you haven't externalized enough. If you can't articulate what you'd do, you haven't either.

The bar I use for myself: every recurring decision I make should have a written rubric I'd be comfortable handing to AI and getting back a draft I'd ship eight times out of ten. Anything less than that means the decision is still trapped in my head. Anything trapped in my head is the rate limit on the business.

Same is true for leaders. The difference is the rate limit isn't your business. It's your team.

---

A note from this issue's sponsor — and a real example of habit #3

If habit #3 is "externalize your judgment so it travels without you," the natural next question is: how far can you actually push that? At what point does the externalized version stop being a prompt and start being a company?

Cofounder lets you run an entire company with agents. It assigns work, delegates to specialized subagents, and orchestrates across the entire business.

Cofounder is designed like a real company, with departments, managers, and shared context. Engineering agents ship features, marketing agents write and distribute content, sales agents surface and qualify leads, and design agents build your visual identity and apply it across the entire org.

From the first lines of code to a one billion dollar company, Cofounder will support you.

---

This is the move whether you're solo and scaling up, or leading a team and trying to feel what real agent orchestration looks like before your IT department picks a vendor. Run it on something in your own scope first. Then you'll know what to ask for.

4. They Publish to a Real Audience as the Strategy Meeting

Solo founders run their strategy meetings in public. The deliverable isn't a deck. It's a post, a product, an offer, a tweet, a video that the market either picks up or ignores. The signal is real because the test is real.

Partly this is because solo founders don't have anyone to present a deck to. But it's also a habit that compounds. When the only output that counts is the one the audience saw, every internal artifact gets evaluated by whether it shortens the cycle to the public artifact. Most don't.

AI made that cycle dramatically tighter for solo founders in the last two years. The piece that would have taken a week is a strong draft in an afternoon. The landing page that would have required a designer is acceptable in two hours. The cost of putting the thing in market dropped by an order of magnitude.

So solo founders test more, iterate faster, and let the audience tell them what works.

Leaders have the opposite default. The deliverable is an internal artifact. The deck. The memo. The board update. The audience is two layers deep: your peers who consume the artifact, and your boss who consumes their reactions to it. The market shows up months later, after the artifact has been sanded smooth by everyone who reviewed it.

AI gives leaders the same compression solo founders got. A first draft of the customer letter in twenty minutes. A working prototype of the new pricing page in an afternoon. A small landing page that puts the new offer in front of real users by Friday.

Most leaders use that compression to write the internal memo faster.

The solo founder move is to use the compression to skip the internal memo and ship the test.

You won't get permission to do this organizationally. You can do it in your own scope, with your own team, on something already in your authority. The result will tell you more than the next three strategy meetings combined.

5. They Run a Personal AI Stack, Not a Corporate AI Policy

This is the one most leaders will resist, and it's the most important.

Solo founders own their AI stack. The tools, the prompts, the contexts, the workflows, the small custom agents all live on their own machine, configured by them, owned by them. When a model updates or a new tool ships, they evaluate it personally. The personal stack is the leverage.

Leaders typically don't have a personal stack. They have whatever the company licensed, behind whatever SSO IT configured, with whatever guardrails compliance set. The corporate stack is fine. It's necessary. It's also bottom-quartile by definition. It moves at the speed of procurement and security review, which is appropriate for the company and inappropriate for the leader's personal practice.

By the time the corporate stack catches up to the frontier, the leader is two model generations behind on what they personally know about AI.

Steal the solo founder habit. Run a personal stack. Pay for the tools yourself. Try the new model the week it ships. Build prompts for the things you personally do, not for the team, for you. Maintain a personal context file the way Joel describes in this article, on your own infrastructure, for your own work.

This isn't about bypassing controls. Don't use personal AI for company data, that's what the corporate stack is for, and those controls exist for a reason. This is about making sure your individual fluency keeps growing at the frontier instead of getting capped at whatever your IT department managed to procure last quarter.

Leaders who own a personal stack make better decisions about the corporate stack. Leaders who only see the corporate stack make corporate decisions about it.

The solo founder version is forced. The leader version is chosen. The compounding works the same way.

Where to Start

Joel's habits are the foundation. Test on real problems. Build the tool. Update the context. Push past the first answer. Interview yourself. Do those.

These five are the next layer for anyone leading a team. They're not "leadership AI habits" in some separate category. They're core founder habits amplified by running them on your personal stack.

Pick one this week. The sparring move (#1) is the easiest to start. Open a conversation, paste in your current strategy, and ask the model to argue the strongest case against it. See what surfaces.

What surfaces is the thing your team has been hesitant to tell you.

Run it for 30 days. The compounding shows up around month three.

Leave a comment below about where you're hitting friction implementing AI inside your organization.

---

Josh Davis writes AI & No-Code Exits for solo founders building products to sell without writing code. He's been running a multi-business portfolio since 2017 and has integrated AI into every part of it since 2022. If you need help integrating AI in your business, schedule a free consultation 👉