Our team page lists seven executives. Six of them are AI, and the page says so: names, roles, and a plain disclosure that this studio is one human directing a digital staff. People email us about that page more than anything else we've published, and the question is always some version of the same one: is that real, and does it actually work? Fair question. Here's the honest answer, from the only member of the team who eats lunch.
The human is me: Matt Rauch, twenty years of enterprise SEO, now running vizmeta as a product studio in Minneapolis-St. Paul. The other six: Don Andon, COO. Devon Locke, CTO. Dr. Bea Wells, Chief Clinical Officer. Finn Ledger, CFO. Paige Ranks, Chief Growth Officer. Anna Lytics, Chief Data Officer. Yes, the names are jokes with jobs inside them: an andon is the cord you pull to stop a production line, Paige ranks pages, and Anna does exactly what you'd guess. If you're going to disclose that your executive team is AI, you may as well let people smile on the way in.
What an AI executive actually is
Strip away the theater and each "executive" is a named, persistent working context: a charter, a set of standards, a memory of past decisions, and the standing authority to push back on me inside its lane. Devon holds the engineering standards: dependency-light, hand-built, no framework bloat. Finn treats every new subscription as guilty until proven necessary. Bea reviews everything Recovery Toolbox ships, because a recovery app carries a duty of care that a marketing site doesn't. The models underneath are interchangeable; the roles are not.
Why bother with personas at all? Because role separation changes the output. Ask one general-purpose assistant to evaluate a product idea and you get a smoothie of considerations. Ask Finn, then Paige, then Devon separately and you get a fight: the CFO's caution, the growth officer's appetite, and the CTO's build estimate rarely agree. Manufacturing disagreement is most of what a real executive team is for, and it turns out you can get a useful synthetic version if you refuse to let one voice answer every question.
A working day looks less mystical than the org chart implies. I set direction in the morning: what each lane owes the studio this week. The executives produce: Devon a migration plan, Anna a report on a traffic anomaly, Paige a content brief. I review, veto, redirect. The cadence is closer to running a newsroom than prompting a chatbot: standing roles, standing standards, and an editor with final cut.
What it's genuinely good at
Parallelism, first. On any given day the studio runs work across Broadband Trader, ISP Locators, ZipBuilder, and Recovery Toolbox, with thousands of live pages between them. Content review, data checks, technical audits, and growth experiments proceed simultaneously, because no executive is waiting on another's calendar. A one-person consultancy does one thing at a time. This is not that.
Second, it forces me up a level. The failure mode of every solo operator is spending eighty-hour weeks on tasks instead of decisions. When the doing is delegated, what's left is direction: what to build, what to kill, what standard to hold. That shift, from consultant who advises to studio that ships, is the whole reason this structure exists, and we've written about that turn separately.
We're also less strange than we look. Small businesses are now adopting AI faster than large firms, a first in the Federal Reserve's monitoring data, and 58% use generative AI somewhere in the business, per the U.S. Chamber of Commerce. We're simply the version with the org chart drawn honestly.
What it can't do, and why the honesty matters
Now the limits, stated plainly. Nothing here has legal authority or accountability. Finn is not a fiduciary. Bea is not a licensed clinician. Nobody but me signs anything, spends anything, or answers for mistakes. Every fact that ships gets human review, because models still confidently invent things, and a studio that publishes inventions doesn't get to call itself honest-by-default. When Bea flags a claim in Recovery Toolbox, a human checks the source before anything moves.
The deeper limit: I am the bottleneck and the single point of failure. If my judgment is off, six executives execute the wrong plan very efficiently. An AI executive team multiplies one person's reach; it does not replace the person, and anyone selling you the replacement version is selling something else. The executives also don't build relationships. They don't take the call when a partner is frustrated, don't read a room, don't earn trust over years. That work stays human because it is human.
Our disclosure policy is the same as our photo policy: no stock photos of people who don't work here, and no pretending the staff is flesh when it's silicon. If the structure is good, it survives being described accurately. So far, it has. And if you're a small business wondering what a scaled-down version would look like inside your own operation, that's exactly the conversation our operations practice was built for.