I've Been Running ClawdBot for Two Weeks. Now I'm Building a Business Around It.

I have squirrels in my attic.

Stay with me.

We were quoted $1,500+ to get them out. But I had a vague memory that when we bought the house, there was some kind of wildlife control warranty buried in the closing docs. I didn't want to spend an hour digging through files to find out.

So I asked my ClawdBot to do it. His name is Tars, and he talks just like the humanity-saver from Interstellar.

He found the document — faster than a full-text search could have — read the terms, and came back with the bad news: I had an obligation to pay a small yearly fee and submit a letter to keep the warranty active. Bummer. But it saved me 30 minutes to an hour of searching, and at least I now know I didn't waste $1500.

That's when it clicked. Not "oh, AI is cool." I've been past that for a while. It clicked that this specific tool is going to change how normal people interact with AI. Not the ChatGPT-style "let me write you a paragraph." The "let me go do something real on your computer while you handle other things" kind.

I've been living in this world for 18 months

I need to back up.

I've been working 100% on building with AI for the past year and a half. It's literally all I do. I'm driving my family nuts with it.

Some context: I bootstrapped and sold my last company for over $100m. I'm golden. I could be at the lake. I could be skiing. I've eaten glass, walked through hell, and come out the other side. Nobody would blame me for coasting.

I'll spend more time in future posts on why I'm not coasting. Short answer: I believe in something. Longer answer coming soon.

But I will say this — I'm doing this phase in what I call "gentleman startup style." I started by hiring engineers, not locking myself in front of a computer 14 hours a day. I've done that. I have nothing left to prove about my work ethic. This time I'm building smarter.

My company, Future Obvious, has been building Force Multiplier AI — a platform to help SMBs become AI-first. Think: organized AI agents that handle real business work, with visibility into what they're doing, audit logging, and a control layer so the business owner isn't flying blind. Also better chat.

We've been heads-down on this since November and are deploying to our first design partners by March. We've already built CollegeCraft.AI, but teenagers are difficult and I'm not made for TikTok.

Then ClawdBot happened.

Two weeks ago, I installed ClawdBot — which became MoltBot, which became OpenClaw (the naming saga is its own lesson in trademark law). I've been running it daily since.

It's insane.

150,000+ GitHub stars. Scientific American, CNBC, IBM, Wired, Forbes — everyone's writing about it. Developers are calling it "AI with hands." People are running it on Mac Minis 24/7 as a personal assistant that never sleeps.

And it's similar to what we were already building with Force Multiplier — except with a totally unhinged disregard for control and security.

Cisco's research team called it "a security nightmare." Palo Alto Networks flagged a "lethal trifecta" of risks. Credentials stored in plaintext. No audit trail. No organizational controls. For personal use on a spare laptop? Incredible. For a business touching customer data and sending emails on your behalf? You need more.

I looked at what we'd already built — the oversight architecture, the visibility layer, the organizational controls — and thought: what if we just wire our control and audit layer on top of this thing?

So that's what we're doing.

I put up moltiplier.ai today. The offer:

We take OpenClaw — the most capable open-source AI agent available — and deploy it for your business with the management layer it desperately needs. Unified visibility. Audit logging. A control layer to organize agents by team and function. Hardened deployment following the best practices that every security researcher is begging people to use.

This isn't a pivot from Force Multiplier — we're still building it. They serve different needs.

Force Multiplier is the structured approach: managed processes with intelligence baked in. Think defined workflows, specific business roles, predictable outputs. It's for the business owner who says "I need an AI that follows my process for qualifying leads" or "I need something that preps meeting briefs the way my best analyst does." Less autonomy, more precision.

OpenClaw is the opposite end of the spectrum. It's "here's a smart agent with access to my tools — go figure it out." Incredibly powerful. Also incredibly unpredictable without a management layer around it.

Most businesses will want both. The structured AI employee for repeatable, high-stakes workflows. The autonomous agent for everything else. Either way, they need visibility and control — which is the layer we've been building all along.

Why I moved fast

The OpenClaw news cycle is peaking right now. The security conversation is happening right now. Every business owner who sees their tech friend's OpenClaw handling their email is going to ask the same question: "Can I get that for my company?"

The answer today is: not safely. Not without the management layer.

So I built the landing page, set up payments, and went live. Not because the product is finished — it's not. But because the fastest way to learn what the market wants is to put something in front of people and see who reaches for their wallet.

A $50 refundable deposit tells me more about real demand than a hundred "yeah, that's cool" conversations ever could.

If you run a business and want to be one of the first to deploy OpenClaw with proper organizational controls, go to moltiplier.ai. If you want priority access, put down the deposit.

I'm documenting the whole journey here. More soon.

— Matt