The Done-With-You Model Explained — What It Is and Why It Works
You have two options for selling your AI expertise right now, and both of them are broken. You can sell a course — film yourself explaining Claude prompting for 8 hours, sell it for $200, and watch 94% of your buyers never finish it. Or you can sell done-for-you services — configure the client's entire AI stack yourself, bill $10K, and become a bottleneck who can't take a vacation because every new client adds hours to your week. The done-with-you model is the third option that sits between these two failure modes, and it works specifically because it refuses to be either of them.
What Done-With-You Actually Means
Done-with-you means the client does the work in real time while you guide. They share their screen. They click the buttons. They write the prompts. They configure the integrations. You sit next to them — virtually, in most cases — and tell them what to do, why it works, and what to do when it breaks. They leave each session with a finished piece of their setup AND the knowledge to maintain it without you. The "with" is load-bearing. It's not done for them. It's not done by them alone using your recorded instructions. It's done in tandem, in real time, with your expertise applied to their specific situation.
The model is not new. Executive coaching has used it for decades. Music lessons use it. Physical therapy uses it. Any field where the client needs to develop a skill — not just receive a deliverable — has some version of this structure. What's new is its application to AI tool setup, where it solves a problem that neither courses nor done-for-you services can touch: the client needs to use the tools after you leave. If they don't understand what was built, your work has a half-life measured in weeks. The first time something breaks or needs adjustment, they're back to square one — or back in your inbox asking for help you didn't price for.
Why Courses Fail for Most Buyers
Course completion rates are brutal. The industry-wide average for online courses sits between 3% and 8%, depending on whose data you trust. [VERIFY] That means for every 100 people who buy your $200 course on AI automation, somewhere between 3 and 8 will actually finish it. The other 92-97 will watch the first module, maybe the second, then life happens and the course becomes another tab they feel guilty about not opening.
The problem is not the content. The problem is the format. Courses are asynchronous and self-paced, which sounds like a feature but functions as a bug. "Self-paced" means "whenever I get around to it," which means "never." There's no accountability, no external structure, and no one watching if you skip a week. The purchase itself provides the dopamine hit — "I'm investing in myself" — and for most buyers, the purchase is the entire experience. The learning was supposed to come later, but later never arrives.
For AI tool setup specifically, courses have an additional problem: they're generic. A course teaches you how to set up n8n in general. It cannot teach you how to set up n8n for your specific business, with your specific tools, your specific data sources, and your specific workflow bottlenecks. The gap between "I understand how this works in theory" and "this is actually running in my business" is where 90% of course buyers get stuck. They need someone to look at their specific situation and say "no, not that node — this one, connected to that."
Why Done-For-You Fails for Most Sellers
Done-for-you sounds like the premium play. You charge $10K-$25K, you build the whole thing, the client gets a polished deliverable, everyone's happy. Except you're now a bottleneck. Every new client adds 20-40 hours of work to your calendar. Your revenue is capped by your available hours. And the ceiling arrives faster than you expect — somewhere around 3-4 concurrent clients for a solo operator, which means you're at capacity the moment things start working.
The deeper problem with done-for-you in the AI space is maintenance. AI tools update constantly. The Claude API changes behavior between model versions. n8n ships new nodes that obsolete your custom workarounds. The automation you built in March is throwing errors by June because three of the APIs it calls have updated their authentication flow. If you built it for them and they don't understand it, every maintenance issue becomes your problem. You've created a dependency, not a solution. The client can't troubleshoot because they weren't there when the decisions were made. They don't know why you chose that node over this one, or what that conditional logic is checking for, or where to look when the webhook stops firing.
Some sellers solve this by offering maintenance retainers — $1K-$2K/month to keep things running. That's a valid business model, but it's not a scalable one for a solo operator. You end up managing a growing portfolio of other people's systems, each with its own quirks and failure modes, and your actual capacity for new work shrinks with every retainer you add.
Where DWY Sits
Done-with-you solves both problems by refusing to absorb either one's failure mode. You're not teaching at scale (so you don't need completion rates to make the math work). You're not building for the client (so you don't become the bottleneck when things break). You're transferring a skill in real time, applied to the client's real situation, with a tangible deliverable produced in each session.
The economics are clean. A typical DWY package for AI tool setup runs 3-5 sessions at $1,000-$1,500 per session — call it $5,000 for a four-session engagement. The client gets their AI workflow configured, tested, and running. They also understand how it works, why you made the choices you made, and what to do when something changes. You get paid at expert rates without taking on ongoing maintenance obligations. The engagement has a defined beginning and end. There is no scope creep because the sessions are fixed. There is no dependency because the skill transferred.
The IKEA effect is real and relevant here. Research consistently shows that people value things they helped build more than things that were handed to them. A client who configured their own AI workflow — with your guidance, at your direction, using your expertise — will maintain it, iterate on it, and advocate for it more aggressively than a client who received a finished deliverable they don't fully understand. The effort is the feature, not the bug. When the client does the clicking, they internalize the logic. When you do the clicking, they internalize the dependency.
Why DWY Works Specifically for AI Tool Setup
AI tool setup is arguably the perfect use case for the DWY model. Three properties make it ideal.
First, the tools require ongoing use. This isn't a logo design where you hand over the PNG and walk away. An AI workflow needs daily interaction — prompting, monitoring, adjusting. If the client doesn't understand how it works, they'll either stop using it (wasting the investment) or use it badly (producing worse results than no AI at all). The skill transfer isn't a nice-to-have. It's the whole point.
Second, the tools are genuinely configurable. There's real expertise involved in choosing between Claude and GPT for a given workflow, in structuring system prompts that actually work, in connecting n8n to the right triggers and outputs, in deciding what to automate and what to leave manual. This isn't "install the app and press go." There are dozens of decisions that require judgment, and those decisions differ for every client. A course can't handle that variation. A template can't handle it. A person sitting with the client, looking at their specific situation, can.
Third, the tools change fast. Whatever you configure today will need adjustment within 90 days — a model update, a pricing change, a new integration, a deprecated API. The client who understands the system can handle those changes. The client who received it as a black box cannot. DWY doesn't just deliver the current setup. It delivers the client's ability to adapt the setup when the inevitable changes arrive. In a market where the leapfrog cycle runs quarterly, that adaptability is the most valuable thing you can transfer.
Who DWY Works For (on the Seller Side)
DWY works best for solo operators and small teams with genuine expertise who want to monetize knowledge without becoming either a content creator or an agency. You don't need a course platform. You don't need a video editor. You don't need a team of implementers. You need a Zoom account, a calendar, and the ability to guide someone through a process you've done many times before.
The skill requirement is real, though. DWY requires you to be good at two things simultaneously: the technical work itself and the act of teaching it in real time. Some people are brilliant builders who can't explain their own decisions. Some people are brilliant teachers who haven't done the work recently enough to guide it at the detail level. DWY needs both — the practitioner who can also articulate the reasoning, adjust to the client's pace, and resist the urge to take over when the client is slow. That's a specific skill combination, and not everyone has it.
If you do have it — if you're the person who already finds yourself walking friends through their Claude setup on FaceTime, already explaining to clients why their automation is over-engineered, already diagnosing other people's AI workflows in your head while they describe them — then DWY is just the productized version of what you're already doing for free. The model doesn't require you to become something new. It requires you to charge for what you already are.
This is part of CustomClanker's Done-With-You series — turning AI skills into client revenue.