The reason focus rituals fail is not that people lack discipline — it is that the transition from not-working to working requires cognitive work that most people skip. You sit down, open your laptop, and immediately start reacting: email, Slack, the task that is loudest, not the one that matters most. The AI prompt at the start of a focus ritual does one specific thing: it forces articulation. You have to name the session type, the energy level, the specific output, and the likely failure mode — and in doing so, you shift from reactive mode into intentional work before the session begins.
The first step of the 4-Minute Gate ritual is an AI prompt. Its purpose is to externalize the context-setting and intention-surfacing that would otherwise happen slowly and inefficiently in your own head.
Each of the prompts below is ready to use now, for a specific session type. Copy the one that fits, fill in the brackets, and run it before your next session.
Prompt 1: For a Writing or Content Session
Use when: You are about to write — a draft, an article, a document, a proposal — and you need to start at the right place rather than from the beginning of the document again.
“I’m starting a writing session — [duration] minutes. I’m working on [piece or project], currently at [specific point: e.g., ‘the middle of section 2’ or ‘draft 1, roughly 60% done’]. My energy today is [low/medium/high]. What’s the most important thing to accomplish in this session, and what’s the writing-adjacent activity I’m most likely to use as a delay mechanism?”
What to do with the response: Take the suggested output, turn it into your Step 3 intention sentence, and use the named delay mechanism as a conscious flag. When you notice yourself drifting toward it, you have already pre-committed to seeing it for what it is.
What good output looks like: A specific output recommendation (e.g., “Complete the argument in section 2 and write the transition into section 3”) and a named delay mechanism that is accurate to your actual patterns — things like “re-reading earlier sections,” “checking references,” or “reformatting headings.” A sign the prompt failed: the AI suggests “avoid distractions” or “stay focused” as the delay mechanism. That is not a named pattern, it is generic advice. If you get that, add: “Based on the session type I described, what is the most common form of productive-feeling procrastination in writing sessions?”
Prompt 2: For a Coding or Engineering Session
Use when: You are picking up a coding session — especially after a break, a meeting, or a day away from the codebase — and need to re-enter the right context rather than the last one you remember.
“I’m starting a coding session — [duration] minutes. Project: [name]. Picking up from [last thing you were working on]. My energy is [low/medium/high]. What’s the most important thing to ship or move forward, and what’s the most likely rabbit hole to avoid?”
What to do with the response: The “rabbit hole” answer is the most valuable part. Rabbit holes in engineering sessions are usually adjacent improvements that feel important but are not on the critical path. Naming one in advance makes it easier to defer rather than follow.
What good output looks like: A concrete next action (“Resume the failing test in the authentication module, specifically the token refresh edge case”) and a specific rabbit hole (“refactoring the helper functions you noticed yesterday — that’s a separate PR”). A sign the prompt failed: you get “focus on your most important task and avoid scope creep.” That is filler. Push back: “What specific technical distraction is most common when picking up [type of work] mid-task?”
Prompt 3: For a Strategic Thinking or Planning Session
Use when: You have time blocked for thinking — not doing — and you want to exit with a decision or recommendation rather than a longer list of things to consider.
“I’m starting a strategy or planning session — [duration] minutes. The key question I’m trying to answer or decision I’m trying to make is: [question]. My current thinking is [brief summary]. My energy is [low/medium/high]. What’s the most important thing to think through, and what would a useful output look like by the end of this session?”
What to do with the response: Strategic sessions fail most often when they produce good thinking but no committed decision. Use the AI’s suggested output to define what a session completion looks like — not “think about X” but “have a written recommendation on X.”
What good output looks like: A specific output definition — something like “a written two-paragraph recommendation with a preferred option and the key trade-off you are accepting.” A sign the prompt failed: the AI suggests outputs like “clarify your thinking” or “explore the decision space.” Those are process descriptions, not completable outputs. The session needs a deliverable, even if that deliverable is a short document only you will read.
Prompt 4: For a Review or Synthesis Session
Use when: You have documents, research, feedback, or data to process — and you need to exit with something extracted rather than something read.
“I’m starting a review session — [duration] minutes. I’m reviewing [documents, research, notes, output] for [purpose]. What I’m trying to extract or decide is: [one sentence]. My energy is [low/medium/high]. What’s the key thing to look for, and what would make this session successful rather than just thorough?”
What to do with the response: The distinction between “successful” and “thorough” is the useful framing here. Thorough review sessions often produce comprehensive reading and no usable output. The AI prompt is designed to surface what you are actually extracting, which keeps the session purposeful rather than comfortable.
What good output looks like: A clear answer to “what would successful look like” that names a specific artifact or decision — “a ranked shortlist of three vendors with the deciding criteria documented” rather than “a good understanding of the options.” A sign the prompt failed: the AI describes a thorough process for reviewing (“read carefully, take notes, identify themes”) without defining what the session should produce. Redirect: “What is the minimum output that would make this session worth doing?”
Prompt 5: For a Low-Energy or Constrained Session
Use when: You have a short window — 20 to 45 minutes — and your energy is not high enough to tackle the hardest thing on your list, but you want to make real progress rather than busywork.
“I have [short duration — 20 to 45 minutes] and my energy is low. I need to make progress on [project or task]. What’s the smallest useful piece of work I can complete in this window — something that actually advances the project, not just busywork? And how should I start it?”
What to do with the response: Low-energy sessions fail when the intention is too ambitious. The ritual step is especially important here because the default under low energy is either avoidance or busywork. This prompt designs a realistic intention from the start rather than discovering halfway through that you aimed too high.
What good output looks like: A task that is smaller than you would normally assign to a session — specific enough that you can confirm you did it. “Write the first two sentences of the introduction” is a valid low-energy task. “Make progress on the introduction” is not. A sign the prompt failed: the AI recommends the same work you would do in a full-energy session, just with a note to “pace yourself.” If that happens, push back: “Assume I can only sustain 60% of my usual output quality. What should I actually do with this time?”
Common Pitfalls
Pitfall: Filling in the brackets with vague information. “I’m working on a project, medium energy” produces a generic response that could have been written without the AI. Fix: be specific about where you are in the work (“draft 1, three sections done, stuck on the argument in section 2”) and honest about energy level. The prompt only surfaces useful information if you give it accurate input.
Pitfall: Spending more than 90 seconds reading the response. The ritual step is one input in a four-minute process. If you spend five minutes analyzing the AI’s output, you have replaced the ritual with a new form of warm-up procrastination. Fix: skim for one useful thing — a priority, a flag, a first action — and move on.
Pitfall: Using the same prompt for every session type. A writing prompt optimized around delay mechanisms is not useful for a synthesis session that needs an output definition. Match the prompt to the session type rather than defaulting to whichever one you used last time.
Pitfall: Skipping the prompt when the session feels obvious. Obvious sessions still benefit from articulation. The most experienced practitioners of focus rituals report that the pre-session prompt often surfaces something they had not consciously noticed — an energy level lower than expected, a dependency they had forgotten, an easier entry point. “I know what I’m doing” and “the ritual is unnecessary” are often the same thought.
How to Use These
Each prompt replaces Step 1 of the 4-Minute Gate. After running it:
- Step 2: Spend 60 seconds on context review (last session endpoint + one reference).
- Step 3: Write your one intention sentence, informed by the AI’s response.
- Step 4: Define your first action and produce it.
Total: 4 minutes.
For the research behind the 4-Minute Gate — including why the ritual works best when it is the same length every session and how to adapt it for remote and fragmented work environments — see the Complete Guide to Building a Focus Ritual with AI.
Your action for today: Pick the prompt that matches your next session type, fill in the brackets now — before you close this tab — and save it somewhere you will see it when you sit down.
Tags: AI prompts, focus ritual, deep work, knowledge work, productivity quick wins
Frequently Asked Questions
-
Do I need to use these prompts exactly as written?
No — adapt them to your actual context. The bracketed fields are placeholders for information you fill in from your real situation. The structure is what matters: you're surfacing your session type, your energy state, your current context, and the one most important output. The exact phrasing is secondary. If a prompt feels mechanical, rewrite it until it feels like a genuine question you'd want answered.
-
How long should I spend on the AI's response?
Sixty seconds or less. The prompt step is one minute of a four-minute ritual. Skim the response for one useful thing — a reminder of the right priority, a flag on a likely distraction, a suggested first action — and move on. If the response is unhelpful, note that and either revise the prompt or skip it in favor of a written self-check. The AI response is an input, not a decision.
Frequently Asked Questions
-
Do I need to use these prompts exactly as written?
No — adapt them to your actual context. The bracketed fields are placeholders for information you fill in from your real situation. The structure is what matters: you're surfacing your session type, your energy state, your current context, and the one most important output. The exact phrasing is secondary. If a prompt feels mechanical, rewrite it until it feels like a genuine question you'd want answered.
-
How long should I spend on the AI's response?
Sixty seconds or less. The prompt step is one minute of a four-minute ritual. Skim the response for one useful thing — a reminder of the right priority, a flag on a likely distraction, a suggested first action — and move on. If the response is unhelpful, note that and either revise the prompt or skip it in favor of a written self-check. The AI response is an input, not a decision.