Motivate Me to Outsource
Written By: Shane Clark on August 4, 2026
Why would someone who outsourced 75,000 hours pause before delegating?
Because AI collapsed the cost of execution. I can start most tasks by voice in about 35 to 40 seconds, so any handoff now has to beat that on speed and quality before I outsource it. Delegation still wins often, but it has to earn the work now.
A project comes in. Before I assign it to anyone, I ask a series of questions. Not whether the work matters, or who is free, or who has done something like it before. I ask one thing first. What is the fastest path to a high quality result, and is a handoff even part of that path anymore.
Here is the honest premise, and it is uncomfortable to say after the last fifteen years. I am not automatically motivated to outsource anymore. The default flipped. It used to be obvious that I should hand work out. Now, more often than not, the burden has moved the other way. If a task is going to leave my hands, something has to motivate me to send it out instead of just doing it myself. So motivate me to outsource. That question sits under every project now.
This is coming from someone who bet on delegation
I am not writing this as someone who dislikes delegation. I have outsourced roughly 75,000 hours over the years (that is about 417 hours a month, every month, for 15 years), most of it to teams in the Philippines I have leaned on hard. I believe in delegation. That is exactly why this shift is worth naming, because it comes from someone who bet on delegation for years, not against it.
When a person who spent that long handing work out starts pausing before he assigns, something structural changed. It did. The tools changed, and the tools reset the math. I have written about the same shift from the other side, from outsourcing to insourcing powered by AI agents.
What the handoff now competes against
Let me make it concrete. A few tasks come in that need to happen on a client’s staging site. The old move is to write up four tickets, assign them, explain the context, and wait. Instead I open my mouth and say one thing.
“Hey Claude, go into Teamwork, it is already logged in, hand out these four tasks, come up with a game plan, and then execute them on the staging site that is already logged in.”
That is the whole handoff. By the time I would have finished explaining the job to a person, the game plan exists and the four tasks are already moving on staging. No tickets, no waiting, no back and forth about who has access to what. The work started while I was still talking, in about 35 to 40 seconds. So that is the clock every handoff now runs against. Under a minute, from my mouth to work in motion.
Delegation always had overhead, now it stands out
Delegation is not free, and it never was. There is a cost to explaining the task, a cost to waiting, a cost to reviewing what comes back, and a cost to closing the gap between what I meant and what I got. We accepted that overhead for years because execution itself was expensive. One person could only produce so much in a day, so spreading the work was the smart play.
AI collapsed the execution side of that equation. The cost of producing a first version of almost anything dropped close to zero. The overhead of delegation did not drop at all. So that overhead, which used to hide inside a bigger number, is now the biggest number in the calculation.
I want to be clear about the goal, because this is easy to misread. The goal is not to delegate everything, and it is not to do everything myself. The goal is to finish high quality work in the least total time. Delegation is one tool for that. It is not the objective.
Every workflow is competing for the task
Once you frame it that way, every project turns into a quiet competition. Me plus AI is one option. An employee plus AI is another. Someone working mostly by hand is a third. Each path has a real speed and a real cost, and the work goes to whichever one reaches a great result fastest.
Competition for the task is not new. A good manager always weighed who should take a job. What is new is the benchmark. The bar for fast and good moved, and it moved a long way in a short time. A workflow that looked efficient two years ago can be the slow option today, even when the person running it is excellent.
So what actually motivates me to outsource
Here is the part that matters if you want my work, or anyone’s work, in this new setup. A few things move the needle, and talent is only one of them.
Access moves it. If you have AI in your hands and know how to use it, you start the race in a different place. Context moves it even more. If you already understand my systems, my documentation, my prompts, my client history, and my procedures, then handing you a task is cheap, because I do not have to rebuild the whole situation first. And a shared, AI-first workflow beats simply being available. Available is not the same as ready. The person who plugs into the same production model I run is worth more to a project than a more talented person running a different one.
That is what motivates me to outsource. Not loyalty, not habit, not guilt about keeping busy hands busy. Fit inside the production model.
Friction is the tax, and the credits question is the tell
Reliability lives in here too, and it shows up as friction. Every point of friction lowers the value of a handoff. Waiting is friction. An unclear process is friction. Inconsistent output is friction. A workflow I cannot predict is friction. Each one is a small tax, and the taxes stack until the handoff is not worth it.
So picture the fork I stand at when a task arrives. Option A is that I spend 35 seconds and my AI starts the task immediately. Option B is that I hand it to my team, and because they are not set up with AI on their side, the work happens manually, or the conversation turns into who is supposed to be paying for the AI in the first place.
That second path is the tell. The moment a handoff turns into who is covering the credits, the decision is already made. Not because the credits cost much. They do not. It is a signal that the person is not set up inside the production model this project runs on, which usually means they are not the right fit for this particular task. And that is okay. It just means the honest answer is simple. I will not assign it to you. I will do it myself.
None of that makes anyone on my team less capable. They work the way they are provisioned to work, and if no one has handed them the tools, manual is their honest best effort. But option A takes under a minute, and option B carries overhead before a single piece of real work begins. Put those two side by side and the choice mostly makes itself.
It is not about the credits, and it is not about talent
Let me protect this from two easy misreadings. It is not really about the price of AI credits. Credits are cheap. This is about removing friction and pushing more finished work out the door. Throughput is the prize.
And it is not about whether someone is talented. I work with talented people. The real question is whether they are operating inside the same production model I am. A brilliant person on a high friction workflow loses to a solid person on a smooth one. That is a strange thing to say out loud, but it is what the numbers show.
Workflow versus workflow
So here is where I have landed, after 75,000 hours of believing the opposite. The people who thrive will be the ones who pair their own expertise with AI and strip the friction out of how they work. They make themselves easy to hand things to. Their context already matches mine. And they run a workflow that plugs into mine instead of fighting it. That is a learnable skill, and it has less to do with raw talent than with how you set yourself up. This is one piece of a larger change in how task work itself is shifting.
The competition is no longer person against person. It is workflow against workflow. The workflow with the least friction and the highest quality wins the work, and it is not close.
This is the operational half of a bigger point I made in Learning AI Is the Easy Part. Using AI is the easy part. Delivering with it is the real skill. So if you want me to outsource, and I do want to, motivate me. Build the workflow that beats 35 seconds and a sentence. If you want help building that kind of workflow, that is exactly what my team and I do. An AI business automation audit maps where the friction lives, and our AI business automation services build the workflow around it. The owners who feel this shift early, and the people who retool for it early, are the ones who spend the next few years quietly pulling ahead.
Frequently asked questions
No. I still outsource, and I have sent out about 75,000 hours of work over fifteen years. What changed is that outsourcing is no longer the default. A task now has to beat the AI path on total time and quality before I hand it out, and plenty of tasks still do.
No. Credits are cheap. The credits question is only a signal. When a handoff turns into who pays for the AI, it usually means the person is not set up inside the same production model, and that friction, not the price, is what kills the delegation.
The opposite. Expertise still matters. The people who thrive will pair their skill with AI and strip friction out of how they work. The competition is not person against person anymore, it is workflow against workflow, and that setup is something anyone can learn.
Fit inside the production model. That means AI access, shared context, my documentation and prompts, and a low friction, reliable workflow. If someone already knows my systems and can plug in, handing them work is cheap and I do it gladly.
About 35 to 40 seconds. I can speak one sentence, send a set of tasks into a logged-in tool, have it build a game plan, and start the work on a staging site before I would have finished briefing a person. That is the clock every handoff runs against.
