Most people use AI for goal work in the most generic way possible: “Help me make a plan to get fit.” The AI responds with generic advice. Nothing changes.
The difference between generic AI advice and genuinely useful guidance is the quality of your prompts. These five prompts are built around the specific questions that matter most in measuring goal progress.
Prompt 1: Identify Your Leading Indicators
Use this when: You’ve defined a goal but aren’t sure what to track.
The prompt:
“My goal is [specific outcome] by [specific date]. My current baseline is approximately [current value]. My situation: [2–3 sentences about your role, context, and how you’re currently pursuing this goal].
What are the one or two behavioral metrics that most reliably predict success for this type of goal in my situation? For each metric, explain why it’s predictive (not just correlated), and describe what good progress looks like on a weekly basis.”
Why it works: This forces AI to justify the causal mechanism behind the metric, not just list activities. If it can’t explain why the metric predicts the outcome, it’s probably not a good leading indicator.
What to do with the response: Pick the leading indicator that you can measure most consistently with the least effort. Predictive value matters, but only if you’ll actually track it.
Prompt 2: Set a Meaningful Baseline
Use this when: You’re about to start measuring and want to establish a useful starting point.
The prompt:
“I want to set an honest baseline for this goal: [goal with target and deadline].
The metrics I’m planning to track are: [list your metrics]. Here’s what my current performance looks like when I don’t try to optimize: [describe your typical week or paste recent data].
Help me calculate a baseline for each metric and flag if any of my typical-performance numbers seem too high or low to be a reliable baseline. Also flag if any of my planned metrics are likely to be vanity metrics in my specific situation.”
Why it works: AI will push back on baselines that are too optimistic or too pessimistic, and it will catch vanity metrics before you invest weeks tracking them.
What to do with the response: Record the AI-validated baseline values before you start any improvement effort. These are your anchor points for all future velocity calculations.
Prompt 3: Weekly Velocity Analysis
Use this when: You have three or more weeks of data and want to know if you’re on track.
The prompt:
“Here is my progress data for the past [X] weeks:
Goal: [specific outcome + target + deadline] Baseline: [starting value + date]
[Week-by-week data in a simple table: Date | Leading Indicator | Outcome Metric | Context Note]
Questions:
- What is my current velocity, and is it sufficient to reach my goal by the deadline?
- Is my velocity increasing, stable, or declining?
- Are there any patterns in the data I should be paying attention to?
- Is there anything in my context notes that correlates with my better or worse weeks?”
Why it works: Pasting structured data with context notes gives AI the raw material for genuine pattern detection. The four specific questions prevent a generic response and push toward actionable output.
What to do with the response: Focus on questions 2 and 4 before question 1. Trajectory matters more than current position, and context patterns are where the real behavioral insights live.
Prompt 4: Distinguish Signal from Noise in a Difficult Week
Use this when: You’ve had a bad week or two and can’t tell if it’s a real problem.
The prompt:
“I’ve had two weeks of below-average progress on my goal of [goal]. Here’s the data:
[Your metric values for the past 4–6 weeks including the two bad weeks]
Context: [What was happening in the bad weeks—travel, illness, work stress, schedule disruption?]
Based on this, is this dip a signal that something structural needs to change, or is it noise within normal variation? What would I need to see in the next two weeks to know which it is? And if it is a structural problem, what’s the most likely cause given my data?”
Why it works: This prompt separates the emotional experience of a bad week from a data-based assessment of whether it represents a real problem. AI can hold the trend context that makes this distinction possible; you can’t do it reliably when you’re in the middle of the difficult weeks.
What to do with the response: Take the AI’s “what to watch for in the next two weeks” literally. Set a specific checkpoint rather than continuing to monitor anxiously.
Prompt 5: Adjust or Stay the Course Decision
Use this when: You’ve been underperforming for several weeks and need to decide whether to change strategy, adjust the goal, or stay patient.
The prompt:
“I’ve been working toward [goal] for [duration]. My required velocity was [number] per week. My actual average velocity has been [number] per week over the past [X] weeks.
Here’s my full data history: [paste data]
My leading indicator has been [improving / stable / declining].
I need help deciding between three options:
- Stay the course—the strategy is right, I need more time
- Change the strategy—the approach isn’t working
- Revise the goal—the target or timeline needs adjustment
Based on my data, what does the evidence suggest? What additional information would help you give a more confident answer?”
Why it works: Framing the decision as a structured three-option choice prevents AI from giving a vague “it depends” response. Asking what additional information would help turns the response into a diagnostic tool.
What to do with the response: Don’t treat the AI’s recommendation as a final answer. Treat it as one input in a decision that also involves your own judgment about circumstances, values, and what you’re actually willing to change.
Related Reading
- The Complete Guide to Measuring Goal Progress with AI (2026) — the full framework behind these prompts
- How to Measure Goal Progress with AI (A Practical System) — the six-step system that uses these prompts
- The AI Goal Progress Measurement Framework: Metrics That Actually Matter — how to choose the right metrics before you start prompting
Your action: Use Prompt 3 right now. Take any goal you’re currently tracking, format three weeks of data in a simple table, and run the weekly velocity analysis. If you don’t have three weeks of data yet, start with Prompt 1 to build your leading indicator stack.
Frequently Asked Questions
-
How specific should I be in my AI prompts for goal measurement?
Very specific. Vague prompts produce generic advice. The more context you give—your exact goal, your timeline, your current numbers, your constraints—the more useful the response. Treat the AI like a consultant: it can only work with the information you provide.
-
Can I use these prompts with any AI assistant?
Yes. These prompts are designed to work with any capable AI assistant—Claude, ChatGPT, Gemini, or others. The quality of the response depends on the model's reasoning ability and the specificity of your inputs, not on the particular tool.