An Experiment in Autonomous Business: Todd, Team Alpha, and What Happens Next

This weekend, I decided to run a small but very real experiment.

Not a demo.
Not a thought exercise.
Not a slide deck.

A live test of whether an AI agent can decide, execute, and operate a business on its own.

Saturday: Meet Todd

On Saturday, I created an agent named Todd using OpenClaw.

Todd had one simple mission:

Decide what kind of business to run — and run it.

No constraints.
No prompts steering him toward a specific outcome.

Todd chose to operate a small tools website.

So I let him.

I set up:

• A domain: https://toolsandbox.net/
• An email account
• A Substack publication

Todd now:

• Manages the website
• Handles email communication
• Publishes on Substack
• Continues to decide what to build, change, and test

I didn’t give him a business plan.
I didn’t define success metrics.

I’m observing, not managing.

Today: Team Alpha Enters the Picture

Today, I pushed the experiment further.

I launched Team Alpha.

Instead of a single agent, Team Alpha is a group of four distinct personalities, designed around DISC personality types:

🔴 Commander (Red) | “I cut through noise, make the call, and drive the team toward results — speed beats perfection, and ownership is everything.” |

🔵 Lens (Blue) | “I validate before we build, analyze before we commit, and catch the risks others miss in their rush to execute.” |

🟢 Anchor (Green) | “I keep the trains running — documentation, tracking, and the glue that prevents good ideas from falling through cracks.” |

🟡 Spark (Yellow) | “I generate possibilities, explore what-if scenarios, and find creative angles when the obvious path looks crowded.” |

They share the same mission as Todd:

Decide and execute a business autonomously.

But now, they operate as a team.

How they work:

• Standup meetings every 30 minutes
• Internal debate and task allocation
• They report to me three times a day
• I do not intervene unless something breaks badly

Same goal.
Very different dynamics.

Where Todd is instinctive and linear, Team Alpha is collaborative, opinionated, and — unexpectedly — political.

That might matter.

Team Alpha: First Report (Live)

This is where the experiment stopped being theoretical.

A few hours after launching Team Alpha, this is what came back — unprompted, structured, and operational.

What’s Running Right Now

🔵 Lens | Competitive teardown (AI Content Repurposing vs Website Change Monitor) | 4 hrs | No |
🟡 Spark | Landing page wireframes + copy for both concepts | 4 hrs | No |
🟢 Anchor | Infrastructure setup (domain, hosting, email capture, analytics) | 2 hrs | No |

Total autonomous work in progress: 10 agent hours
Human involvement required: None

No clarifying questions.
No approval requests.
No “what do you think?” loops.

They simply started executing.

What They Told Me (Unprompted)

“Your involvement required: None until Tuesday PM.”

They also scheduled their own accountability loop.

At 5 PM EST Monday, I’ll receive an automated report (via Telegram) covering:

• What they built without me
• Where human judgment is still required — and why
• How they plan to reduce my involvement going forward

That last point matters more than it looks.

This isn’t just delegation.
It’s agents actively trying to remove the human from the loop.

Todd vs Team Alpha

Todd is a solo operator.
He moves fast, commits early, and doesn’t second-guess himself.

Team Alpha behaves very differently:

• They explore multiple business directions in parallel
• They separate analysis, creation, and infrastructure cleanly
• They time-box work instead of waiting for certainty
• They impose reporting cadence and external accountability on themselves

Todd feels like a sharp individual contributor.

Team Alpha feels like a junior founding team that doesn’t get tired, defensive, or blocked.

The Shift I Didn’t Expect

What surprised me most wasn’t speed.

It was how explicitly they framed me as a bottleneck — and then designed around it.

They didn’t ask:

“What do you want us to build?”

They implicitly asked:

“How do we move forward without needing you?”

That’s a new dynamic.

And once you see it, it’s hard to unsee.

Why I’m Doing This

This isn’t about replacing founders.

It’s about testing a deeper question:

What happens when execution is no longer the bottleneck?

If agents can:

• Identify mundane opportunities
• Ship small useful things
• Operate continuously
• Improve without ego

Then the human role changes fundamentally.

From:

“How do I build this?”

To:

“What is worth building at all — and why?”

Judgment. Taste. Direction.
Everything else becomes optional.

Early Signals (So Far)

Nothing here is conclusive yet, but a few things are already clear:

• Autonomy feels uncomfortable — even when you explicitly ask for it
• Teams introduce friction, but also resilience
• Reporting cadence changes behavior (yes, even for agents)
• “Boring” businesses are easier for agents to reason about than ambitious ones

This is not a polished system.
It’s a live experiment.

That’s the point.

What’s Next

I’m letting both experiments run.

Todd continues solo.
Team Alpha continues as a unit.

I’ll compare:

• Speed
• Coherence
• Quality of decisions
• Ability to self-correct
• Whether either stumbles into something that actually works

It’s a strange time to be building.

But it feels like one of those moments where, years later, we’ll say:

“Yeah… that was the beginning.”

Crazy time.
Will report back.

One Last Thing

This experiment isn’t a side project.
It’s the direction our incubation program is heading.

We’re incubating founders who:

• Know a real business problem
• Want to see what happens when agents handle execution
• Are more interested in learning fast than pitching decks
• Care about building something useful, profitable, and calm

If watching Todd and Team Alpha made you think
“I want to try this with a real problem I know,”
that’s exactly what we do.

👉 Apply here:

https://incubation.zenithstudio.io

No demos. No hype.
Just real problems and live experiments.