How Do You Plan Your Entire Week in 10 Minutes With AI?

You plan your entire week in ten minutes by talking instead of typing: you open an AI assistant, talk through everything you need to accomplish, and let it build the planner and deliver it before you walk into the office on Monday. I'm Christian Saab, founder of CPG Recruitment, and I do this every Sunday morning while I feed my son. Ten minutes of talking out loud, and the plan is in my inbox at 8 AM Monday. Since March 2020 I've built teams across 25+ industries in Canada and the US while running a second company, and the only reason both stayed upright is that I stopped trying to carry the week around in my head.

What does a 10-minute weekly planning session actually look like?

It looks like a conversation, not a document. Sunday morning, I feed my son, I open Claude, and I talk through everything I need to accomplish that week — the roles we're filling, the clients I owe a call, the decisions that have been sitting on my desk since Wednesday. I don't organize any of it while I'm talking. That's the whole trick. The assistant builds the planner out of what I said and emails it to me Monday at 8 AM, so I walk into the office with my week already mapped.

Ten minutes is not a productivity gimmick. It's the amount of time it takes to say out loud what you already know you have to do. The hour people usually lose to planning isn't thinking time — it's formatting time, and that's the part you should be handing off.

Why does talking to an AI beat writing your own to-do list?

Because talking empties your head and a list only captures what you can already name. When you sit down with a blank page, you write the six things you were already worried about. When you talk for ten minutes, the half-formed items come out too — the supplier who hasn't replied, the interview you keep rescheduling, the thing your operations lead flagged on Thursday that you nodded at and forgot.

I left a six-figure engineering job at 24 to build companies, and engineering is where I learned to look at a week the way you look at a machine: inputs, process, outputs. Your brain is a terrible place to store inputs. It's an excellent place to make decisions. Planning by voice separates the two — you supply the inputs, the assistant handles the process, and you spend Monday on the decisions.

What do you have to tell the AI before it can build a week worth following?

Tell it outcomes, constraints and non-negotiables — not tasks. A task list in, a task list out; that's a reformatting exercise. Give it the three outcomes that would make the week a win, the meetings that cannot move, the hours you refuse to give up, and who you're waiting on. Then it can sequence the week instead of alphabetizing it.

The first session takes longer than ten minutes, because that's where you describe the business: what you do, who your customers are, what a good week looks like, what you personally should never be doing. Once that context exists, every Sunday after it is a ten-minute conversation.

Why should the plan land Monday morning instead of Sunday night?

Because a plan you read on Sunday night gets re-litigated all evening, and a plan you read at 8 AM Monday gets executed. Sunday is for your family. If the planner arrives Sunday at 9 PM, you will open it, find one thing you don't like, and spend the rest of the night working in your head instead of resting.

Timing the delivery is a small detail that does real work. The plan exists before you walk in, and you meet it at the moment you can actually act on it. That's the difference between a document and a system — and systems are what let a business run without the owner holding it together, which I've written about in the three parts of a business that runs without you.

How do you keep a weekly plan tied to the two engines of your business?

Force every week to account for both engines: the sales engine that brings customers in, and the talent engine that puts the right people in place to serve them. If your plan is seven days of operations and firefighting, you're maintaining a business, not growing one. I make the assistant ask the same two questions every week — what moved the sales engine, and what moved the talent engine.

That discipline is what produced 28 new clients in six months against an industry average of 24 to 25 a year. Those weren't the weeks I worked hardest. They were the weeks the right conversations were already on the calendar before Monday started. If you want the longer version of the framework, read what the two engines every business needs to grow actually are.

What does this look like when you're running more than one company?

It stops being a nice habit and becomes the only thing holding the week together. I run CPG Recruitment, founded in March 2020, and Booked55, the B2B sales system I built and proved inside CPG before selling it as software. Two companies means two sets of priorities competing for the same Monday, and the loser is always whichever one didn't get written down.

Planning by voice makes that competition visible. When both engines are in the same ten-minute conversation, you can see the week where recruitment gets four days and sales gets none — before you live it. A search doesn't go from 68 days to 37 because somebody hustled. It goes from 68 to 37 because every step had an owner and a date before the week started.

What makes AI weekly planning useless?

Three things: no context, no review, and treating the output as sacred. Paste a raw task list into a chat window with no background and you'll get a prettier task list. Never tell it what actually happened last week and it will keep planning for a business that doesn't exist anymore. And if you treat Monday's plan as a contract instead of a starting position, you'll abandon the whole practice the first Wednesday that goes sideways.

The plan is a draft of your week, not a promise. When Wednesday blows up — and it will — you don't throw out the system. You spend two minutes telling it what changed and let it re-sequence what's left.

Where should you start if you've never done this?

Start with one Sunday and one ten-minute conversation, and don't try to automate anything yet. Talk through the week, read what comes back, and notice what it got wrong — that gap is exactly the context you forgot to give it. Do that three weeks in a row and you'll have a planning system that's genuinely yours.

And if the week you keep planning is full of interviews you don't have time to run and roles you can't fill, the planning isn't the problem. That's a talent engine problem, and it's the one my team fixes for a living.

If your weeks keep getting eaten by hiring, talk to us: CPG Recruitment builds recruitment processes for owners across 25+ industries, and you can book a conversation with me directly. If it's the sales side of the week that's empty, Booked55 is the system I built for exactly that.

Frequently Asked Questions

How long does the first AI planning session take?

Longer than ten minutes — usually closer to an hour, because the first session is where you describe your business, your role and what a good week looks like. That context is a one-time cost. Every Sunday after it is a ten-minute conversation, because the assistant already knows what it's planning around.

Does this replace a project management tool?

No. A project management tool tracks work across weeks and people; this decides what you personally do with the next five days. They solve different problems, and the plan should point at the tool rather than duplicate it.

Does it have to be Sunday?

No. What matters is that there's a gap between planning the week and starting it, so the plan is ready before the first meeting rather than written during it. Pick the day that gives you that gap and keep it.

What do you do when the week falls apart by Wednesday?

Re-run it. Spend two minutes saying what changed and what's still outstanding, and let the assistant re-sequence the back half of the week. The point of a ten-minute system is that redoing it costs almost nothing.