An empty afternoon isn't free time
A general AI assistant can reach a lot of your life now — memory that carries across chats, your documents in a project, a whole Cowork workspace, even your calendar and your inbox. It can see the meeting that moved and the reply you owe, and it can put something in a free slot.
What it can’t see is your side of it: the work you meant to do, what that work will actually cost you in hours, and where your time really went last week. So it finds the gap on Thursday and offers it to you — nothing has told it that Thursday is already spoken for.
Connected to MakeTimeFlow, it stops guessing at the gaps. It can ask what your day honestly holds, work from the direction you set when you were thinking clearly, and leave what the two of you decide somewhere it’s still true next week.
Your calendar and your inbox are mostly other people’s claims on your time. MakeTimeFlow is your own side of that ledger — what you intended, what it costs, and what it was for. An assistant that can read both will tell you a day is overfull instead of quietly filling it, and what you settle on lands in the same place your weekly review will look for it.
The method it plugs into
Left to itself, an assistant will improvise a system for you, and the one it improvises is almost always there’s a gap, do more. Connecting to MakeTimeFlow puts it inside a considered one instead — built on decades of research into goals, attention, and performance, and built just as deliberately against the ways that research goes wrong.
- Attention gets planned, not just time. Blocking an hour doesn’t produce an hour of work. What produces it is naming what done looks like and deciding in advance what you’ll do when the thing that always derails you derails you. So that’s what gets asked for before a timer starts, rather than after you’ve lost the morning.
- What you’re working toward is a story, not a checklist. You write success in the present tense across five years, this quarter, and this week. A vivid narrative is something you can steer by; a list of objectives is something you tick. Because the assistant reads that story, it can tell you why a task earned your Thursday.
- The known failure modes are designed against. Hard, specific goals raise performance — and reliably produce tunnel vision, stubborn loyalty to a bet that stopped paying, and a life optimised on one axis. So the framework underneath spreads attention across six dimensions of flourishing rather than one, treats a Smart Bet as a bounded experiment you’re allowed to abandon, and opens the weekly review on what you delivered rather than what you missed. A review that starts with your failures is one you quietly stop doing.
You keep the pen throughout. Nothing is written without your yes, and writing a success story never activates it — committing stays a deliberate act. The aim was never to hand your judgment to an assistant. It’s that one working inside a method you control, on what’s true this minute, is far more use than one improvising from your calendar.
A conversation that ends when you close the window is advice, however much it could see. One that changes what’s in front of you tomorrow, and leaves the reason behind, is something you can actually build a life on.
Start here
- Connect Claude and AI Agents — what a connected assistant can do, how to connect it, and what you’re approving.
Going further in Claude Code and Cowork
A connection gives an assistant your live work. If you use Claude Code or Cowork, the flow plugin also gives it the method — how to get you out from under a pile, how to plan a day it will tell you honestly is too full, how to start work you’ll actually finish, and how to end a week knowing where your time went.
- Getting Started with the flow Plugin — install it in two lines, then the four commands that make a first day.
- flow Plugin Reference — every command, the rhythm they’re built for, and the traps worth knowing.