A calendar that defends itself: an AI scheduler instead of your willpower

A focus block in your calendar survives until the first meeting someone schedules on top of it. Then you either move it by hand — or, more often, you just sacrifice it. Willpower runs out; a calendar doesn't. And because defending your time is ninety percent mechanical work (spotting the collision, finding a replacement slot, moving it, letting people know), it can be handed off to a routine.

The end state is simple: your calendar has a written constitution, and every morning a check runs to see whether anyone broke it. When someone schedules a meeting inside your deep-work block, you don't find out the moment a notification pops up — your morning summary already has a line like “the Analysis block on Wednesday at 9:00 collided with a meeting, suggesting a move to Thursday at 8:30, everything else holds.” You approve it or you don't. The decision stays yours; hunting for a slot and recalculating your week doesn't.

This guide moves through six phases: writing the rules in plain language, connecting Google Calendar with a connector, a daily collision watch with the complete routine brief, defending against calendar erosion, prepping briefing notes before meetings, and a weekly planning ritual. One rule governs everything: AI proposes, the human approves. The routine may work inside your own calendar and must never touch anyone else's — moving a meeting is a social act, not a technical operation.

A typical scenario

Tomáš, a lead analyst, dutifully blocks three mornings every Monday for deep work. By Wednesday, only one is left — colleagues click into the open slot, he accepts the meeting, cancels the block, and tells himself he'll move it later. He doesn't. By Friday he realizes he spent the whole week reacting and made no real progress on anything. Lunch got skipped four times, and he showed up to two external meetings unprepared, because he'd planned to read the briefing “on the way there.”

After setting up the guard, his week looks different. Monday morning, his inbox has a summary: fifteen hours of deep work spread across three mornings, two collisions with meetings he'd already accepted, and a proposed move for each, with a reason why that particular slot. Wednesday morning, the routine tells him someone scheduled a meeting inside his Analysis block; Tomáš accepts the meeting and confirms the block's move to Thursday with one tap. The day before an external meeting, he finds a briefing note waiting in the Drafts folder: what it's about, what was agreed last time, three questions to ask.

After two months, the numbers look different: out of fifteen planned hours of deep work, twelve actually happen instead of five, lunch gets skipped twice a month instead of four times a week, and meeting prep costs him five minutes instead of thirty. Willpower didn't do that — a rule that someone remembers every day at seven in the morning did.

Phase 1: the calendar's constitution, written in plain language

Most people who try an AI scheduler and end up disappointed skip exactly this step. A tool with no rules just reshuffles blocks into whatever's open — and an open slot in your calendar isn't the same thing as a good slot. The difference between “that hour was free” and “that hour was good” is your own preferences, and nowhere are they written down.

Three layers instead of two

A calendar is better described in three layers, not a split between “free” and “busy”:

• Untouchable. Never moves and nothing ever gets scheduled into it: lunch, picking up the kids, the end of the workday, commutes, a standing commitment outside work. There are few of these, and that's exactly why they hold.
• Protected. High priority, but allowed to move — typically deep-work blocks, meeting prep, a weekly review. The key word is move, not cancel: a protected block that gets bumped has to reappear somewhere else the same week.
• Open. The rest of your day, where meetings are allowed. If there isn't much of this layer left, the problem isn't the scheduler — it's how much you've committed to.

A constitution built from your real calendar

Writing rules from memory is worse than deriving them from what actually happened — your calendar from the last two months is a more honest source than your mental image of your own week.

Connect to my calendar and go through the last [8 weeks]. Don't
change or create anything — this run is read-only.

Analyze what my week actually looks like:
1. How many hours a week meetings take up, split into internal
   and external. How that differs across days of the week.
2. Which parts of the day have the most uninterrupted time with
   no meetings, and when the day is most fragmented.
3. How long the gaps between meetings are — how many are
   shorter than 30 minutes (practically unusable).
4. What repeats every week in the calendar and what's one-off.
5. When I most often start and end, and how often I have
   something scheduled after [5:30 p.m.] or over lunch.

At the end, suggest five rules that would follow from that
data — phrase them as sentences I could hand to a scheduler as
an instruction. Where you're not sure because the data doesn't
support it, say so instead of guessing.

You'll get back a picture that usually differs from how it feels in two ways: you have less uninterrupted time than you think, and short gaps between meetings do the most damage to it. Read point 3 first — if most of your gaps are under thirty minutes, no amount of planning will get you deep work until the meetings cluster together. Treat the suggested rules at the end as a starting point, not a finished result; the model is guessing from event titles, and those often don't reflect what you actually did.

Writing rules a machine and you both understand

Rules don't have to be technical. They have to be unambiguous (not “mornings work better for me,” but “schedule deep-work blocks between 8:00 and 12:00”), decidable (you have to be able to tell what to do on a collision from them), and short — one screen, not a manual.

Here's the output from analyzing my calendar and my own notes
on how I'd like the week to look:
[paste analysis and notes]

Turn that into a rules card for my calendar scheduler.
Structure:

1. UNTOUCHABLE — events and times that never move and never
   get anything scheduled into them. For each, give the exact
   time and what to do if someone sends an invite into it.
2. PROTECTED — blocks that take priority over a regular meeting
   but are allowed to move. For each: how many hours a week,
   which part of the day, minimum uninterrupted length, and
   by when a replacement slot has to be found.
3. PRIORITIES — the order that decides a collision. Write it as
   a ranked list, not a wishlist.
4. DAY RULES — earliest start, latest end, maximum meetings per
   day, required break between meetings, days with no meetings.
5. EXCEPTIONS — who and what takes priority even over a
   protected block.

Write sentences in the imperative, so they can be used as
instructions. For each rule, add what the scheduler should do
when it can't be followed — no rule may be left without a
procedure for the conflict case.

That last paragraph is the heart of the whole card. A rule with no procedure for conflicts is just a wish, and the scheduler will ignore it in practice: without instructions on which is worse — shortening the block, moving it to another day, or pushing it over lunch — it decides on its own, and you're left figuring out why your lunch disappeared. Save the card into a Project (persistent context), so the routine doesn't need it re-supplied every time.

Untouchable blocks that actually hold

Three sentences worth having in the card verbatim, because they cover most erosion cases:

• Lunch. “Nothing gets scheduled between 12:00 and 1:00 p.m., not even internal meetings. Don't accept an invite into this time — decline and suggest two alternatives.” Lunch is the first thing that disappears from a calendar, and the last thing you notice.
• End of the workday. “Nothing gets scheduled after [5:30 p.m.]. If someone sends an invite for later, list it in the summary and suggest a morning slot instead.” Without this rule, the day stretches slowly and quietly.
• Deep work. “At least [12] hours a week in blocks of at least 90 minutes, preferably between 8:00 and noon. A block may be moved, never canceled; the replacement slot has to be in the same week.” Ninety minutes is the floor below which a block just turns into a longer context switch — the tip on calendar blocks for deep work explains why.

Phase 2: connecting Google Calendar with a connector

The goal is for AI to see your calendar directly and be able to create and move your own events in it; describing your week to a chatbot by hand lasts about three days.

What the connector will see, and what to allow it

Connectors (built on the MCP protocol) are a standardized way to link AI to everyday tools — calendar, email, files, notes. In settings, you connect Google Calendar, sign in with your account, and approve a scope of permissions; from that point on, the model can read events, search for open slots, and create or edit events in your calendar. The tip on connectors as USB-C for AI covers the concept.

Think through four things before you turn it on.

Which account this runs on. A calendar contains meeting titles, other people's names, and sometimes the content of a negotiation. It belongs on a work account with contractual data protections, not a free chat. Give sensitive meetings a neutral name in the calendar — “meeting” instead of names and numbers.

Which calendars to connect. Your work one, yes. Your personal one too, but ideally only as busy/free information — the scheduler needs to know you're unavailable Tuesday at five, not why. In Google Calendar, sharing with “free/busy only” visibility is built for exactly this.

What it's allowed to write. Creating and moving your own blocks, yes. Canceling other people's meetings, responding to invitations on your behalf, or sending invitations to other people, no — and it's not enough to set the permission; this ban needs to be written into the routine's brief too.

The calendar as the single source of truth. If some of your commitments only live in your head, the scheduler doesn't know about them, and its suggestions will feel out of touch. Before you turn it on, get the recurring commitments you “didn't bother writing down” into the calendar too.

Dry run: a week's audit with no changes

Before the routine starts moving anything, let it go through one week read-only. You'll find out whether it sees the right calendars, and whether your rules actually work in practice.

Go through my calendar for next week and compare it against the
rules card in this Project. Don't change, create, or cancel
anything.

Return:
1. Which of my rules are broken next week — for each, the
   specific event, day and time, and which rule it breaks.
2. How many hours of deep work I have scheduled versus how many
   the card requires.
3. Days with no pause between meetings, and places where a
   15-minute one would be enough.
4. Events whose title doesn't tell you which layer they belong
   to (untouchable / protected / open) — list these separately
   so I can fill in a rule.
5. How much of next week is already booked and how much is
   still open.

For each finding, write what you'd suggest doing, but don't do
anything. Where you're missing information to decide, say which
piece is missing.

Point 4 is a goldmine: every event that doesn't fit a layer is a missing sentence in your card. Point 5 is your reality — if next week is eighty percent booked, no amount of scheduling will manufacture deep work, and it's time to deal with the number of meetings, not their arrangement.

Phase 3: the daily collision watch

You have the rules, you have the calendar connected. Now turn it into a scheduled task that runs every morning.

Anatomy of the brief

The guard's brief has six parts: scope (which calendars and how far ahead), checking (what to compare against the rules), resolving collisions (what order to try options in), what it may do on its own (and what it may not), what to leave to me, and a summary. The second-to-last point matters most: a guard that only flags problems is an alarm clock; a guard that silently reshuffles half your week is a liability. The useful middle ground: it only acts on its own in your own calendar, on things that are easy to undo, and proposes everything else.

Complete brief for the daily watch

Every business day at [7:00 a.m.], check my calendar and watch
over the rules from the card in this Project.

1. SCOPE
Go through today and the next [10 business days] across all my
connected calendars. Treat events from my personal calendar as
busy/free information only — don't look at the content.

2. CHECKING
Find every place where reality diverges from the card:
- a meeting scheduled into untouchable time
- a meeting that overlaps a protected block
- a protected block that disappeared or got shortened
- a day with no pause between meetings
- a meeting outside working hours
- a week where the required deep-work hours don't add up

3. RESOLVING COLLISIONS
For each collision, propose a fix in this order:
a) move the protected block elsewhere the same day
b) move it to another day in the same week, same part of the day
c) split it into two pieces, but never below [90] minutes
d) shorten it, by at most [30] minutes
e) if none of that works, flag it and write what would need to
   free up — offer two concrete meetings to cancel or delegate
For each suggestion, write ONE SENTENCE on why that particular
slot.

4. WHAT YOU MAY DO ON YOUR OWN
You may move or create MY OWN blocks that have no other
attendees, and add [15] minutes after an external meeting for
me to write notes.
YOU MAY NOT: cancel or move meetings with other attendees,
respond to invitations on my behalf, invite anyone, change the
titles or descriptions of other people's events, or delete
anything.

5. WHAT TO LEAVE TO ME
List for a decision: collisions where the fix would mean moving
a meeting with other people; invitations awaiting a response;
new recurring events I don't know about.

6. SUMMARY
Write me a short overview, one screen at most:
- what you changed and why (a list with before and after times)
- what you're proposing and waiting on my approval for
- deep-work status: how many hours are scheduled this week
  against the card's target
- what you were missing from the rules to decide
Sort by urgency, today first.

Three places where the brief most often breaks down. Recognizing “my own” blocks — the model works off attendees, so a block you once accidentally invited a colleague to stops counting as yours; check what it actually moved the first week. Time zones for external meetings: if you work with someone abroad, spell out in the card which zone the check runs in. And the order of options in point 3 — without it, the scheduler reaches for shortening the block, because it's the easiest option, and you end up with three forty-minute fragments instead of one solid one.

On top of the daily watch, it's worth considering the whole suite of email-and-calendar routines described in the tip on routines over mail and calendar.

Two-minute approval

Read the summary in the order it arrived; skim through the changes already made. Slow down at the “awaiting approval” section: every item has three legitimate outcomes — approve, decide differently (and write why — that's the seed of a new rule), or leave it because you know something the routine can't.

One thing to always do: check that the moved block still makes sense. A deep-work block moved to Friday at 4:00 p.m. technically satisfies the rule, but it's not going to happen. If a suggestion like that comes up twice, the card is missing a sentence about when you're actually still capable of focused work.

Phase 4: defending against calendar erosion

Erosion isn't one big meeting that eats your day. It's a week where fifteen hours of deep work becomes five without you ever deciding that's how it should be. Each individual concession is reasonable — the meeting really is necessary, the colleague really does need to talk — but the sum is a disaster, and nobody's responsible for it.

Erosion has to be measured first

Until you have numbers, it's a dispute of feelings. With numbers, it's a fact you can take to your manager.

Run a calendar-erosion audit for the last [8 weeks]. Compare
what I had scheduled with what actually happened.

Return:
1. How many hours of deep work I blocked per week and how many
   stayed untouched. A table by week, plus the trend.
2. What most often replaced the blocks: split into event type
   (internal meeting, external meeting, one-off urgent matter,
   personal), with the share of each type.
3. The most frequent source of collisions — which recurring
   meetings or which people most often land in protected time.
   Don't interpret names, just list the frequency.
4. Which days and hours erosion is strongest.
5. How many times lunch got skipped and how many times I had
   something scheduled after the end of the workday.
6. Three systemic changes that, based on this data, would help
   the most — ranked by impact, each with an estimate of how
   many hours a week it would recover.

Base this only on what's in the calendar. Where you can't tell
whether an event actually happened, say so instead of guessing.

You'll get back a table that's harder to stay quiet about. Point 3 is the most useful: erosion is often half caused by two or three recurring meetings that can be moved with a single conversation. Read point 1 knowing the calendar doesn't know what you actually did — if a block stayed in the calendar while you spent it answering emails, the statistics won't catch that. Which is why it's more honest to look at the trend than the absolute numbers.

Four mechanisms that actually slow erosion

Booking windows. Instead of letting anyone pick any open slot, only let meetings into selected parts of the day — say, Tuesday and Thursday afternoons. Google Calendar and booking tools both support this; the effect is immediate, because the rest of the week stops existing as an option.

Buffers. Fifteen minutes after every external meeting to write notes, and five minutes between internal ones. Without them, the day fuses into one long stretch, and notes get pushed to the evening.

A no-meeting day. One day a week where nothing gets scheduled. It works better than scattered blocks because it's easy to communicate — “I'm not available on Wednesdays” doesn't need an explanation.

The swap rule. When a meeting lands inside a protected block, the block has to reappear somewhere else the same week before you accept the meeting. Of these four, it's the only rule that's entirely about you — and also the only one the routine can enforce automatically.

A polite “no,” prepared in advance

Most erosion happens because in the moment you don't have a decent reply ready. Have one prepared in advance.

Someone's asking me for a meeting at a time that's protected
according to my card. Details: who [role and relationship],
what it's about [topic], proposed time [time], how urgent it is
[estimate].

Prepare three draft response options:
A) I offer two other slots from the open layer, no explanation
B) I suggest a shorter format (15 minutes, a call, in writing)
   and explain what we both gain from that
C) I give way and go straight to saying where I'm moving the
   protected block

Rules: no more than 4 sentences, no apologizing for having other
plans, no explaining my system, no lying about already having a
meeting. Pull the times from my calendar, so they're genuinely
open and inside a booking window.

At the end, write in one sentence which option you'd pick and
why — and whether there's a situation where giving way is the
right call.

Option B tends to be the most useful and the least used: a large share of requests for an hour-long meeting can be handled with fifteen minutes or three sentences. The prompt's last paragraph deserves attention — sometimes the right answer is “give way,” and a planning system that doesn't allow for that is one you'll eventually break yourself.

Phase 5: prepping briefing notes before meetings

The calendar guard doesn't only have to defend your time. Since it can already see what's coming, it can also prep for it — and shrink the number of meetings that run long only because people are trying to remember things.

A pre-meeting brief

In [2 days] I have a meeting, [event title from the calendar],
with [who, role, company]. Prepare a one-page briefing note.

Use: the calendar event including its description and
attachments, my correspondence with these people, and notes
from the last meeting if you can find any.

Structure:
1. What it's about and why we're meeting (3 sentences)
2. What we agreed last time and which of it is done — for each
   point, note where you got it from
3. Open items where they're waiting on me, and where I'm
   waiting on them
4. Three questions I should ask so the meeting produces a result
5. The decision I want the meeting to land on
6. What I should bring or send ahead of time

Don't make anything up. Where you don't have a source, write
[MISSING: what to look up] instead of a plausible-sounding
sentence. One page maximum.

Point 2 is the reason this pays off: most meetings start with ten minutes of trying to remember who promised what. The [MISSING] marker is mandatory — without it, the model will fill in “we agreed on a deadline by end of month” convincingly enough that you'll repeat it in the meeting. The tip on preparing for a meeting with AI covers the deeper process.

The “tomorrow” routine

The brief is worth automating: the evening before, so you're not scrambling in the morning.

Every business day at [4:30 p.m.], look at what I have
tomorrow on my calendar and prep me for it.

For every event tomorrow with other attendees:
- one sentence on what it's about and with whom
- what we last discussed (from mail and notes)
- what they'll want from me at this meeting
- whether I need to send or prepare anything yet today
For external meetings, also add a short brief using my template
and save it as a draft email to myself.

Then check tomorrow as a whole:
- do I have pauses between meetings per the rules card?
- is there anywhere with more than [3] hours of meetings in a
  row?
- do I still have at least one block for focused work tomorrow?
Where the answer is no, suggest a fix, but don't make it.

At the end, write three things I have to get done tomorrow, and
one I can postpone without consequences.

The last paragraph is unexpectedly valuable: the model sees both your calendar and your inbox, and its guess about what to postpone is at least unbiased. Treat it as a suggestion to weigh — priorities are your call, and the model doesn't know office politics or what you promised someone verbally.

Phase 6: the weekly planning ritual

The daily watch handles collisions. But once a week, someone has to decide what the week should look like — that's work you don't delegate, it can just be set up so it takes twenty minutes instead of an hour.

When and how

Friday afternoon or Sunday evening, twenty minutes, always the same. The ritual has three steps: close out (what didn't happen last week, and what to do about it), decide (what matters next week), and place (where it goes on the calendar). AI handles the first and third steps; the middle one is yours.

Plan my next week. Inputs: my calendar for next week, the rules
card in this Project, and this list of things I need to fit in:

[paste 5-8 items, each with a time estimate and a deadline if
it has one]

Work through it like this:
1. List what's already fixed on the calendar, and how much open
   time is left after that — in hours, by day.
2. Check whether my items actually fit into the remaining time.
   If not, say so directly and suggest what to postpone —
   don't shrink the estimates just to make it fit.
3. Place the items into specific blocks. Respect: untouchable
   times, a minimum of [90] minutes for deep work, mornings for
   demanding items, no two demanding blocks back to back
   without a pause.
4. For each block, write in one sentence why it's placed there.
5. Leave [20] percent of the week empty as a buffer for whatever
   comes up.
6. At the end, write three risks: what's most likely to break
   this plan, and what to do when it happens.

Output it as an overview by day. Don't create anything yet —
wait for my approval.

Point 2 is the whole point of this prompt: most weekly plans fail because you schedule forty hours of work into thirty open hours — and AI will happily let you, unless you explicitly forbid it. Guard point 5 just as fiercely: a week with no buffer is a plan that falls apart by Wednesday. The tip on timeboxing picks up methodically from here, covering how to work with blocks once they're already on the calendar.

Once you approve it, let the blocks get created — and keep an eye on one thing: name blocks specifically. “Deep work” gets sacrificed easily; “Q3 budget draft” much less so, because the title itself makes clear what you'd be giving up.

Retrospective: plan versus reality

Once a month it's worth comparing what you planned against what actually happened.

Compare my weekly plans for the last [4 weeks] against what
actually stayed on the calendar.

1. How many planned blocks happened unchanged, how many got
   moved, and how many disappeared entirely.
2. For the ones that disappeared: what replaced them, and could
   it have been predicted?
3. Systematic mistakes in my planning: am I underestimating
   time? scheduling more than fits? putting demanding work at
   the wrong time of day?
4. Which types of work regularly go missing from my calendar
   (for example, prep, review, admin).
5. Two changes to the rules card that, based on this data, would
   help the most.

Be critical. I don't want to hear that it went well — I want to
know where the plan and reality diverge systematically.

Point 3 tends to be uncomfortable reading, and that's exactly why it's worth doing. A typical finding is that the estimates are fine and the buffer is missing — or that demanding work systematically ends up on Friday afternoon.

Boundaries that don't move

Other people's calendars stay other people's. The routine may create and move your own blocks. Canceling a meeting with other people, responding to an invitation on your behalf, or inviting someone — only you get to do that. Moving someone else's meeting is a relationship matter, and an automation has no way of knowing what that person is dealing with.

Nothing gets deleted. Moving, yes; canceling, no. A deleted event is harder to bring back than a moved one, and for recurring events, a deletion can easily spread further than you intended.

Data follows company rules. A calendar reveals who you're dealing with — connect it through a work account with contractual data protections, in line with your employer's policies. Give sensitive meetings a neutral name.

Common mistakes

• Deploying a scheduler with no written rules. The tool then reshuffles blocks based on what's open, and you end up with deep work on Friday at 4:00 p.m. An open slot isn't the same thing as a good slot.
• Making everything untouchable. When everything's protected, nothing is. The untouchable layer should be small — three to five items — so the rest has room to give way.
• Letting the routine touch other people's meetings. Automatically moving someone else's event, or responding to an invitation on your behalf, damages a relationship faster than it saves you time. AI proposes, the human approves.
• Not measuring erosion. Without numbers, it's a dispute of feelings you'll lose. A monthly calendar audit is the cheapest argument you can have.
• Planning a week with no buffer. Forty hours of work doesn't belong in thirty open hours. Twenty percent empty isn't laziness — it's the condition for a plan to survive Wednesday.
• Naming blocks generically. You'll sacrifice “deep work” without a second thought. A specific title — exactly what that block produces — holds up much better.

The best tools

• Claude with a calendar connector — you write your priority and untouchable-block rules as plain sentences; a scheduled task watches for collisions every morning, suggests moves, and preps briefing notes for tomorrow's meetings; the rules card lives in a Project, so you edit it in one place.
• Built-in features of Google Calendar and Microsoft 365 — goals, focus time, booking pages, and automatic declines outside working hours. No extra tool, no extra data sharing; start here.
• Reclaim — a scheduler built around habits and priorities; it reshuffles blocks on conflicts and watches over recurring needs, including breaks.
• Motion — combines a calendar with a task manager and schedules tasks into open slots on its own, based on due dates.
• Booking tools like Calendly — for external meetings; instead of an email back-and-forth, the other side picks from windows you've pre-approved, which also happens to be the single most effective defense against erosion.

What you get out of it

• Time: two to three hours a week just from manual rescheduling and back-and-forth over dates. But the bigger win is in the hours of focused work that would otherwise never happen at all — for people who'd only been promising themselves blocks until now, the difference tends to run into several hours a week.
• Peace of mind: you no longer have to be the guardian of your own calendar; when a block falls, the system suggests where to put it back.
• Quality of work: uninterrupted ninety-minute blocks are dramatically more productive than a chopped-up day, and prepping before a meeting shortens the meeting itself.
• A healthier routine: lunch and end-of-day boundaries hold, because a rule enforces them, not your willpower at the end of a hard day.

Pro tip

Advanced move: have the guard calculate a reactive-to-proactive time ratio once a week — how many hours went to meetings someone else called versus work you scheduled yourself. One sentence in the weekly summary is enough. The absolute number isn't that interesting, but its trend is the best early warning a calendar can give you: when the reactive share climbs three weeks running, erosion has started, and it can still be stopped with one conversation instead of a job change.

And the rule that sits above everything else: the scheduler manages time, not decisions. It may find a slot, recalculate the week, and remind you of what you promised yourself. It may not decide what's important, and it may not touch anyone else's calendar. The moment you move that boundary, you stop gaining time and start dealing with misunderstandings.