Prompts That Work With What You Already Have
You do not need a perfect logging system to start using AI for time awareness. If you can produce a rough list of last week’s tasks with estimated and actual durations—even pulled from memory for the most recent week—the prompts below will generate useful output.
These five prompts are ordered from lowest to highest data requirements. Start with prompt 1 regardless of where you are in your calibration practice.
Prompt 1: The Initial Diagnosis
When to use it: After any week where you have both estimates and actuals recorded, even roughly.
The prompt:
“Here is my task log from this week. Each row has a task name, task category, my estimate before starting, and the actual time it took: [paste your log]. Calculate my estimate-to-actual ratio for each task category. Which three categories have the lowest ratios (meaning I underestimated most)? Are there any tasks where I significantly overestimated?”
What you get: A ranked list of your worst estimation categories, with ratios that make the distortion visible rather than vague. Most people discover the category they underestimate most is not the one they expected.
Run this every two weeks as long as you are in the diagnostic phase.
Prompt 2: Building Your Multiplier Table
When to use it: After two or more weeks of logged data across consistent task categories.
The prompt:
“Based on [number] weeks of task log data, here are my cumulative estimate-to-actual ratios by category: [paste ratios or full log]. Convert these into practical adjustment multipliers rounded to the nearest 0.25. Flag any category with fewer than five data points as ‘preliminary.’ Generate a simple table I can use during planning.”
What you get: A ready-to-use multiplier table derived from your actual data, with uncertainty flags where the data is thin. This table goes into your planning document and gets consulted before you finalize any week’s schedule.
Update this monthly or whenever a category ratio shifts by more than 0.15.
Prompt 3: Contextual Pattern Check
When to use it: After three or more weeks of logging that includes time-of-day and energy-state fields.
The prompt:
“Here is my time log with task category, estimate, actual, time of day (morning/afternoon), and energy level (high/medium/low). Are my estimates more accurate in the morning or afternoon? Is there a correlation between low energy and higher underestimation? Which task categories are most sensitive to energy state?”
What you get: Context modifiers to layer on top of your task-type multipliers. If your afternoon estimates are 25% less accurate than your morning estimates, you need a context modifier—not just a category multiplier—to build an honest afternoon schedule.
Run this once in your second or third month of logging, then revisit quarterly.
Prompt 4: The Pre-Task Scope Check
When to use it: Before starting any complex or unfamiliar task, regardless of whether you have a full log history.
The prompt:
“I am planning to [describe task clearly]. My initial estimate is [X hours/minutes]. What components of this task am I likely forgetting to account for? What typically causes tasks like this to run over? What dependencies should I check before I start?”
What you get: A structured list of the scope elements your optimistic brain skipped. This is the inside view / outside view correction in action: you bring your initial estimate, the AI asks the questions your planning bias suppresses.
Use this on-demand for high-stakes or unfamiliar tasks, and any time your estimate is more than two hours for a single session.
Prompt 5: The Weekly Recalibration
When to use it: Once per week, after your logging practice is established.
The prompt:
“Here is this week’s task log: [paste log]. My current multiplier table is: [paste table]. For each category, calculate this week’s ratio and compare it to my table’s current multiplier. Which multipliers should be updated? What is the trend direction over the last four weeks—are any categories improving or degrading? Suggest any revised multipliers.”
What you get: An updated multiplier table grounded in rolling data, with trend flags that tell you whether a category is stable, improving, or drifting worse. The trend information is as important as the current ratio—it tells you whether your calibration is working.
Run this every week once your logging practice is consistent.
What These Prompts Cannot Do
These prompts are pattern analysis tools, not substitutes for real-time logging. They will only be as accurate as the data you bring to them.
They also cannot account for scope changes mid-task, external dependencies that materialize unexpectedly, or the first instance of a genuinely novel task type. Use a 2x multiplier for anything without a clear historical analogue and add it to your reference library once you have a data point.
For the full framework behind why these prompts work—and the underlying research on time perception distortion—the complete guide to time perception and productivity covers both in depth.
Tags: AI prompts, time awareness, time estimation, planning fallacy, productivity
Frequently Asked Questions
-
Do I need a formal time log to use these prompts?
For prompts 1, 4, and 5, even a rough list of last week's tasks with estimated and actual times will work. For prompts 2 and 3, which build on pattern analysis, two or more weeks of data produce meaningfully better results. -
Which AI tools work best for these prompts?
Any capable large language model (Claude, ChatGPT, Gemini) handles these prompts well. The key is the quality of the data you provide—clear labels, consistent categories, and real-time actuals rather than reconstructed estimates. -
What if I do not have time estimates recorded—only actuals?
You can still use prompts 1 and 5 by asking the AI to identify which task categories take the most time unexpectedly relative to their complexity. Without estimates, however, you cannot calculate your distortion ratio—prompts 2 and 3 require both estimate and actual columns.