Procrastination is not a time management problem. Fuschia Sirois, Tim Pychyl, and others in the field are consistent on this point: chronic delay is driven by emotion regulation failure — specifically, the short-term impulse to avoid a task that generates discomfort (anxiety, boredom, self-doubt, resentment) overriding the longer-term intention to complete it. Advice about to-do lists and productivity systems misses this. The problem is not that people do not know what they should do. It is that something about the task itself triggers an avoidance response, and that response is faster and more automatic than the plan.
What AI can do in this context is act as an externalizing tool. The research on affect labeling — naming an emotion rather than experiencing it — shows that precise labeling reduces the emotional intensity of a response. When you describe your avoidance to an AI in specific terms, you are doing affect labeling in writing. The AI’s follow-up questions push you toward precision you might not reach alone.
The five prompts below address five distinct procrastination situations. They are not interchangeable. Start with the one that fits your current state.
Prompt 1: The Feeling Excavator
When to use: You know you’re avoiding something but can’t name why.
“I’ve been avoiding [describe the task] for [X days]. I keep meaning to do it and then doing something else instead. I don’t think it’s about time — I think something about the task itself feels uncomfortable. Help me figure out what that discomfort is. Ask me a few questions if you need more context about what the task involves and what’s at stake.”
Why this works: Tim Pychyl’s research shows that procrastination is driven by specific negative emotions — anxiety, boredom, self-doubt, resentment, overwhelm. Naming the feeling precisely reduces its authority. The AI’s follow-up questions often surface things you haven’t articulated to yourself yet.
What good output looks like: The AI should ask you clarifying questions rather than immediately diagnosing the problem. A useful first response might be: “When you say you keep meaning to do it and then doing something else — what does ‘something else’ tend to be? Is it random, or is there a pattern?” If the AI immediately names your emotion (“it sounds like you’re feeling anxious about this”), it has moved too fast. The diagnosis should emerge from your answers, not the AI’s first-pass interpretation.
Prompt 2: The Self-Compassion Reset
When to use: You’ve been berating yourself about procrastinating and the self-criticism is making things worse.
“I’ve been putting off [task] and I feel genuinely bad about it. I’ve been telling myself [describe your inner critic’s main message]. I know self-criticism isn’t helping, but I also don’t want to just let myself off the hook. Help me respond to this situation the way a reasonable, kind person would respond to a friend in the same situation — acknowledging the difficulty without excusing the avoidance.”
Why this works: Fuschia Sirois’s research shows that self-compassionate responses to procrastination predict less future avoidance. This prompt explicitly asks for the balance between kindness and accountability that people struggle to find on their own.
What good output looks like: A response that holds two things simultaneously — genuine acknowledgment of the difficulty, and a clear refusal to excuse the avoidance. The test: does the response make it easier to avoid the task, or does it make it slightly harder? If the AI’s response leaves you feeling understood but still comfortable not doing the work, it has overdone the compassion and underdelivered the accountability. Ask it to try again: “More honest about the cost of continued delay, without being harsh.”
Prompt 3: The Minimum Starting Point
When to use: You understand the feeling and have addressed it — now you just need to begin.
“I need to work on [task]. I’m not trying to finish it or do it brilliantly. I just need to begin. What is the single smallest thing I could do in the next five minutes that counts as having started? Be specific about what I should physically do, and help me phrase it as an implementation intention — when, where, and what exactly.”
Why this works: Peter Gollwitzer’s implementation intention research consistently shows that specific if-then plans substantially increase follow-through compared to general intentions. The “counts as having started” framing matters — you’re lowering the psychological threshold, not trying to produce the finished product.
What good output looks like: One concrete physical action — open a specific file, write one sentence, send one message — phrased as an implementation intention: “When I sit down at my desk at 2pm, I will open the document and write the first sentence of the introduction.” If the AI returns a list of steps or a mini-plan, it has missed the point. The ask is for the single smallest starting action, not a project plan. Push back: “Give me one thing, phrased as an if-then.”
Prompt 4: The Catastrophe Audit
When to use: Anxiety about outcomes is the main driver of your avoidance.
“I’ve been avoiding [task] and I think part of it is that I’m anxious about how it will be received. My actual fear seems to be [describe the fear — e.g., ‘that my analysis will be wrong and people will think I’m not competent’]. Help me audit this fear: Is this outcome realistic? What actually happens if this worst case comes true? Am I confusing task execution with something larger?”
Why this works: A lot of procrastination anxiety is evaluative — fear of judgment, fear of failure, fear of what completing the task implies. Externalizing the fear and examining it reduces the threat signal. The AI doesn’t eliminate the risk, but it often reveals that the catastrophized version is much worse than the realistic version.
What good output looks like: Three distinct questions — Is this realistic? What actually happens if it occurs? What larger identity claim is riding on this task? — answered specifically for your fear, not generically about fear of failure. A sign the prompt has succeeded: the feared outcome looks smaller and more recoverable after the audit than before you ran it. A sign it failed: the response reassures you that everything will be fine. That is not an audit; it is comfort.
Prompt 5: The Weekly Procrastination Debrief
When to use: End of week reflection to identify patterns before they compound.
“This week I planned to do [list the avoided tasks] but didn’t get to them. For each one, help me identify: (1) whether this was genuine reprioritization or avoidance, and (2) if avoidance, what feeling was likely driving it. Then help me set one specific implementation intention for each avoided task for next week.”
Why this works: Pattern recognition is powerful. Reviewing avoidance weekly rather than only at crisis point interrupts the compounding effect — where each avoided week makes the task feel bigger and more threatening. The debrief also separates legitimate prioritization from avoidance, which many people conflate.
What good output looks like: For each avoided task, a clear verdict on whether it was genuine reprioritization (the task was legitimately displaced by something more important) or avoidance (it stayed on the list and generated guilt). The implementation intention for each avoided task should be specific enough to schedule: “On Tuesday at 10am, I will open the draft and spend 25 minutes on the outline.” If the AI labels everything as genuine reprioritization, it is being too charitable. Push back: “Assume at least half of these were avoidance. Which ones, and why?”
Common Pitfalls
Pitfall: Using the prompts as a substitute for doing the work. Spending forty minutes exploring your feelings about a task is still forty minutes of not doing the task. These prompts are diagnostic tools, not the deliverable. Fix: set a timer before you run any prompt. When the timer ends, the AI conversation ends. Whatever the output is, act on it immediately.
Pitfall: Describing the task vaguely in the prompt. “I’ve been avoiding a big work project” produces generic output. “I’ve been avoiding writing the risk section of the Q3 board deck, which is due Friday and which I know will require me to disclose a metric my manager doesn’t know about yet” produces useful output. Fix: name the specific task, the specific deadline, and the specific thing that makes it uncomfortable — before you ask any question.
Pitfall: Running Prompt 3 before addressing the underlying emotion. If you have not identified what you are avoiding and why, asking for the minimum starting point produces a list of small tasks that you will also avoid. Fix: run Prompt 1 first. Prompt 3 is for after you understand the avoidance, not instead of understanding it.
Pitfall: Using the weekly debrief as confession rather than analysis. The debrief works when it produces specific implementation intentions for the following week. If you finish the debrief feeling absolved but without any concrete plans, it has become a ritual rather than a tool. Fix: the debrief is complete only when every avoided task has an if-then plan attached.
For the full framework behind these prompts, see the Emotion-First Reset. For the research grounding, see the complete guide to the psychology of procrastination.
Your Action for Today
Pick Prompt 1. Open any AI tool. Fill in the brackets with one specific task you’ve been avoiding. Read the AI’s response and answer its follow-up questions honestly.
You don’t have to do the task today. You just have to name what’s happening. That’s enough.
Frequently Asked Questions
-
Do these prompts work with any AI tool?
Yes. These prompts work with any conversational AI — Claude, ChatGPT, Gemini, or others. The quality of the response depends partly on how much context you give in the prompt. The more specific you are about the actual task and the actual feeling, the more useful the response will be.
-
What if the AI's response isn't helpful?
Iterate. Tell the AI what wasn't useful: 'That was too generic — I need something more specific to my situation, which is...' Most AI tools respond well to that kind of calibration. You can also ask it to take a different angle: 'Instead of helping me plan, help me figure out what I'm actually afraid of with this task.'
Frequently Asked Questions
-
Do these prompts work with any AI tool?
Yes. These prompts work with any conversational AI — Claude, ChatGPT, Gemini, or others. The quality of the response depends partly on how much context you give in the prompt. The more specific you are about the actual task and the actual feeling, the more useful the response will be.
-
What if the AI's response isn't helpful?
Iterate. Tell the AI what wasn't useful: 'That was too generic — I need something more specific to my situation, which is...' Most AI tools respond well to that kind of calibration. You can also ask it to take a different angle: 'Instead of helping me plan, help me figure out what I'm actually afraid of with this task.'