Principles
What I believe about planning.
Five principles. The reasons behind them. These are how I decide what goes into Axis — and what stays out.
Most product principles are post-hoc. The product gets built, the marketing team writes a values page, the values page gets ignored. The principles below aren’t that. They’re the reason Axis is shaped the way it is — and the reason a few obvious-seeming features will never ship.
Each one is paired with the product behavior it produced. If a feature breaks one of these principles, the principle wins. If a shortcut would be faster but introduces friction the user has to maintain, the slower path wins. If the AI could do something for the user but doing it would weaken them, we don’t do it.
If any of this rings true, the rest of the site is worth your time. If not, fair — the product won’t be for you, and that’s its own useful information.
— Ross R., Founder
01
Plans are hypotheses.
A plan is a hypothesis about how the next stretch of time will go. It’s based on what you know now. New information arrives by Tuesday afternoon. A plan that doesn’t change in response to new information isn’t a plan — it’s a wish.
Most planning tools treat the original schedule as the contract. When reality diverges, you’re the one who failed. You drag the block. You feel slightly worse about yourself. You move on.
Axis treats divergence as data. The cancelled deep-work block is telling you something specific: the meeting before it ran long, you under-slept, the project actually doesn’t need that work this week, or you’re avoiding it for a reason worth examining. The trends dashboard surfaces patterns of divergence that no individual instance reveals. The lifecycle log captures every cancellation, every reschedule, every block that ran long.
None of it is meant to make you feel bad. It’s there because the gap between plan and reality is where the useful learning lives. The right tool isn’t one that pretends the plan was right. It’s one that asks what your week is actually telling you.
02
Agency, not automation.
A planning tool can do almost anything for you now. Take your goals. Take your calendar. Take your chronotype. Output a fully filled week. The temptation to ship that is real.
The reason Axis doesn’t is straightforward: a tool that makes all your choices for you doesn’t help you get better at making them yourself. Automation that removes the user from decisions also removes the practice that built the user’s judgment in the first place. The week looks fine. The user gets weaker.
So the AI in Axis suggests — never decides. Every change runs through a pending-changes panel where you accept or reject each item individually. Nothing auto-applies. The AI cannot reach into your calendar without your explicit confirmation. The constitution baked into every request says it plainly: the user always has final authority. The AI advises, never commands.
Yes, this is more friction than full automation. The friction is the point. It’s the moment where you confirm what you actually want, instead of being drifted into a week you didn’t consciously choose. A planning tool should make you better at planning, not less practiced at it.
The deeper version of the argument: the goal isn’t a perfectly optimized week. The goal is a user who, over time, gets sharper at noticing what their weeks are actually about. That doesn’t happen if the system is doing the noticing for them.
03
Completion compounds.
Half-finished work has a specific cost most planners ignore.
Each unfinished thing is a small open claim on attention. Even when you’re not consciously thinking about it, it’s there: a faint background tax on every other decision. The drag scales — more open loops, more cognitive overhead spent maintaining the inventory of what you might get back to.
The cost compounds. The next thing you start arrives with all the prior unfinished things still attached. So you start fewer things. The ones you do start are smaller. Ambition gradually concedes to overhead.
A planning tool should help you finish things — not just start them. That’s the whole game. The slow accrual of completed things is what generates capacity for the next bigger ambition.
Axis is biased toward visibility on what isn’t finishing. The trends dashboard surfaces goal effort versus original estimate. The weekly review explicitly asks what carried over and what closed. Goals carry a "what happens if this slips" field — because most goals don’t have one, and most goals slip. The conveyor belt of tasks-to-blocks is designed to keep moving, not to grow.
This isn’t about hustle. It’s the opposite. It’s about giving yourself less reason to dread what’s already on your list, so the next thing isn’t arriving into a backlog of guilt.
04
Frictionless.
If using the system feels like work, you’ll stop using it.
Productivity tools have a maintenance cost. Every checkbox to flip, every status to update, every screen to navigate is a small tax that accumulates over a session. When that tax gets high enough, the tool stops being net-positive. People log in, see all the things needing maintenance, and close the tab.
If a planner takes meaningful effort to maintain, it will lose to the planner that doesn’t — every time. The bar is high and unforgiving. Auto-save instead of save buttons. Smart defaults that work the first time. Drag-and-drop that doesn’t fight you. Every interaction shaved by a beat or two. Inputs you can fill out in five seconds, not five minutes.
Frictionless is one of the most expensive principles to honor. It looks invisible — and that’s the point. A 500ms debounced auto-save throughout. The dark interface that doesn’t fatigue your eyes on long sessions. The keyboard shortcuts that are actually fast. The week templates that fill in eighty percent of the schedule before you touch a thing. Each of those is a tiny battle, fought at the edge of the implementation, that adds up to a tool you don’t notice while you’re using it.
This is why "frictionless" isn’t a UX detail. It’s an existence-of-the-product detail. A planner that doesn’t do this isn’t a worse planner. It’s a planner you’ll quit. The category is littered with tools that solved the wrong problem brilliantly.
The reverse holds too. The reason Axis tracks habits as binary checks instead of streaks, the reason rituals auto-schedule themselves, the reason session feedback is two clicks instead of a form — all of it is the same principle. Make the maintenance cheaper than not maintaining. That’s how you stay used.
05
Research-driven.
Decisions in the product aren’t opinions. They’re hypotheses I can defend.
The first version of Axis came from something drawn on paper — lines of work blocked out by hour, recovery blocks, personal blocks. It worked because pen and paper forced a structure that matched something true about how the day actually unfolds.
When that became software, the question shifted. Hundreds of small decisions get encoded into the product. How long is a deep-work block? How much recovery do I auto-insert? What’s the energy cost of meetings? What’s a sustainable weekly cap on focused work? How wide is normal chronotype variation? Each of those small numbers, multiplied across every user, would shape the system. Making them on instinct wasn’t going to hold up.
So the product is built on the literature where the literature is solid — and is honest where it isn’t. Some findings hold up well: context-switching costs, recovery biology, behavioral architecture. Some don’t: ego depletion is the headline failure of the replication crisis, and the parole-study citation that gets quoted everywhere is contested. Where research is contested or shows wide individual variation, Axis builds in flexibility instead of locking in a number. Where it’s silent, the defaults are stated explicitly and exposed as adjustable.
The research foundation page on this site walks through which studies hold up, which ones don’t, and what we do about it. The failures aren’t hidden.
The principle ends with this. Research is the starting point. Your behavior is the ground truth. The system gets sharper the longer you use it because what you actually do is more reliable than what any one study says you should.
In closing
These five aren’t aspirational. They’re operational.
When a feature looks good in isolation but breaks one of these principles, the principle wins. When a shortcut would be faster but introduces friction the user has to maintain, the slower path wins. When the AI could do something for the user but doing it would weaken them, we don’t do it.
You can hold a product accountable to its stated principles. That’s the test — whether the principles are real, or whether they’re marketing. The product pages on this site are designed so you can check.
Plans are hypotheses. Agency, not automation. Completion compounds. Frictionless. Research-driven. That’s what Axis is built on.
See how it’s built into the product.
The principles aren’t kept on a values page. They show up in feature decisions — every time.