When we set up Claude AI agents for your business, we’re on your computer for a live session, configuring your machine and getting your agents running on real work. That part is always hands-on. There’s no version where you hand us the keys and walk away.
What comes after is a choice, and it’s the same choice we give clients on pilots and retainers who haven’t done a setup session at all. Once the agents exist, who runs them week to week?
We don’t do outreach. We never come to you. So by the time you’re on a call with us, you’ve already decided something, and I’d rather you decide this part too, with the tradeoffs in front of you instead of discovering them three weeks in.
First, the garage
Most SEO and digital agencies are set up like used car dealerships. There’s a guy out front with slick hair telling you everything you want to hear, because he wants the commission. In the back there’s a mechanic with grease on his hands, working on cars, definitely not presentable to the client. The sales guy is usually honest and means well. But he has never looked under the hood of a car in his life, so he makes promises he can’t keep and hires people he has no way of evaluating.
Before you buy a used car, wouldn’t you want a mechanic to inspect it?
We’re the mechanics. That part isn’t the choice. The choice is whether you hand us the keys, or stand in the garage next to us.
Both paths run on the same engine
Whichever way you go, the ongoing work runs on a weekly MAA cycle. Every Friday, roughly five minutes once you’re in rhythm.
I’ve watched this for years across hundreds of businesses, and the pattern never changes. People either fall off a cliff, missing week after week, or they’re crushing it. I don’t see anyone in between. It’s like the gym. Not a one-time superhero effort, just consistency over a couple of months.
So the real question underneath both options isn’t about technical skill. It’s who holds the wrench on Friday.
The three rules, which never change
You own the data. You own the assets. You supply the raw content.
These are the same rules we publish for every agency we work with, and they apply no matter which path you take. You should own your own marketing whether you work with us or anyone else, because the moment you can’t log in and see for yourself, you’re flying blind.
That third rule matters more than people expect. We don’t generate synthetic content out of thin air, because Google can tell the difference and eventually catches you. We start with your real, raw ingredients: your videos, your Zoom calls, your podcast appearances, your customer stories. What the agents do is amplify the authentic signal of you actually showing how you do what you say you do, in the city you say you do it. That signal has to come from you, and no path changes that.
Option 1: We drive
You’re focused on operating your business and you don’t want to fiddle under the hood. That describes most successful founders and there’s nothing wrong with it.
The agents still live on your side. We operate them and deliver your MAA report every week. You read it, ask questions, approve direction. It feels like an agency engagement, except there’s no black box. The agents are on your infrastructure, you can see what they’re doing, and you can take the whole thing in-house whenever you want. Unlike the agency charging $2,000 a month where nothing has happened in nine months, you can actually look.
The upside is obvious. Minimal time from you, no learning curve, fastest path to output. We’ve done this hundreds of times, so we’re not learning on your account.
The cost is real too. Your team builds no internal capability, so six months in you still can’t run this yourself. Feedback loops are slower because we have to ask you for context instead of you supplying it live. And your output is capped by how much of your business you actually hand us. The agents get smarter the more they know about you and your company, so if that knowledge stays in your head, there’s a ceiling on what they can do. If you decide to bring it in-house in a year, there’s a ramp.
There’s one thing you have to accept for this to work. The more urgent your situation, the more important it is that we coordinate smartly rather than frantically. Imagine an ER surgeon with a patient bleeding out. The family is yelling nonstop, believing this increases urgency and makes the surgeon work harder. And the surgeon, not wanting to appear uncaring, has to stop and reassure them, which takes precious time away from what he knows he needs to do.
If you hire an expert, decide whether you trust them. Option 1 only works if you do.
This is the right call for operators with no spare hours, businesses where the founder is already the bottleneck, and teams that want results before they want expertise.
Option 2: You drive, we ride along
You want to be in the weeds. You’re already nosy about this stuff.
You run the agents on your machine and we supervise while you learn. You’re copied on the working threads in Basecamp so you see how the work actually happens, not a summary of it. And you come to weekly Office Hours, which is a live MAA review, not generalized Q&A. You bring your actual numbers, your actual analysis, and your actual proposed actions, and we tear into them in front of everyone. That’s the whole value. You’ll learn more watching five businesses get their analysis picked apart than from any course.
At premium tiers, $3,500 a month retainers and above, you can bring other people from your team and do private one-on-ones with me or others on our side.
If you bought the setup session, this is the natural continuation. You were already at the keyboard. Option 2 just means you stay there.
What you get out of it is durable capability. In six months you can run and extend this without us. Feedback loops are tight because you correct the agents in real time using knowledge only you have. The output is structurally better, because analysis requires business context, and you have it in a way we never will. Your team can maintain and extend the system without being hostage to us or anyone else.
But it takes real time every single week. Office Hours, thread participation, your Friday MAA. There’s a learning curve and your progress is entirely a function of your engagement. Momentum dies if you go quiet, and this is where the cliff comes in. Two missed Fridays and most people never come back. Someone on your side has to own it, because if it’s everyone’s job, it’s nobody’s.
I’ll tell on myself here. I’ve caught myself running agents in Claude, Codex, and Perplexity simultaneously, spending half my time chasing the latest updates instead of having the discipline to spend eighty percent of it on implementation. That’s me, and this is what I do for a living.
So be honest about whether the tinkering will pull you off your actual business. For some people it’s genuinely fun and it compounds. For others it’s an expensive hobby that quietly replaces the work.
If you’re not going to do MAA every Friday, pick Option 1. That’s not a sales tactic. Option 2 without weekly consistency produces worse results than Option 1, and you’ll have paid in time as well as money.
This is the right call for founders and teams who want to own the capability long-term, businesses with deep proprietary knowledge worth teaching an agent, and anyone planning to scale this across more of the company.
How to decide
One question. Do you want output, or do you want capability?
If it’s output and your calendar is already full, take Option 1. If you want your team running this without us in six months, take Option 2, and commit the Fridays, because that’s the only thing that makes it work.
But if you want the output of Option 1 while resenting the oversight, or the capability of Option 2 without showing up on Fridays, neither path will work. Better to know that now than four weeks in.
You can start on Option 1 and move to Option 2 later. Everything we build is portable and documented, so switching costs you nothing.