These five prompts cover the full arc of building and maintaining a connected goal system — from initial stack audit to weekly review. Copy, adapt the bracketed sections to your situation, and run them.
Prompt 1: Stack Audit — Where Are My Goals Really Living?
Use this when you’re diagnosing a fragmented system or starting from scratch.
“I want to understand how well my current tools support my goals. Here are the tools I use regularly: [list your apps — e.g., Notion, Todoist, Google Calendar, Toggl, Slack]. Here are my current active goals: [list 3-5 goals with brief descriptions]. For each goal, I’d like you to: (1) identify where in my stack each goal ‘lives’ if it’s anywhere, (2) flag where data about progress on that goal is being captured (or not), and (3) identify the biggest gap — the connection that doesn’t exist but should. Ask me any clarifying questions before you start.”
This produces a gap map you can act on immediately. It often surfaces orphan goals — goals that have no home in any tool — and reveals which satellites are completely disconnected from any goal structure.
Prompt 2: SSoT Design — Help Me Build the Right Goal Database
Use this when choosing or redesigning your Single Source of Truth.
“I’m building a goal database to serve as my Single Source of Truth. I use [your task manager] for tasks, [your calendar app] for scheduling, and [your AI assistant] for weekly reviews. My goals are organized as [OKRs / quarterly goals / annual themes — describe your method]. Please help me design a database schema for my SSoT. I need it to: (1) hold all active goals with clear success criteria, (2) track progress in a structured way that an AI can read and analyze, (3) support a weekly update log with dates, and (4) include fields that map to how my task manager and calendar already categorize work. Suggest field names and types. Then tell me which field is the most important one most people forget to include.”
The answer to that last question is almost always the update log with date stamps — and the prompt surfaces it in a way that makes it memorable.
Prompt 3: Automation Design — What Should I Actually Automate?
Use this before building any Zapier or Make flows.
“Before I build automations to connect my tools, I want to make sure I’m automating the right things. Here is my current workflow for weekly goal review: [describe how you currently pull data, update progress, and run your review — be specific]. Here are the tools involved: [list tools]. Where in this workflow am I doing repetitive, low-value data-gathering that could be automated? And where am I doing thinking work that should stay manual? Give me a prioritized list of three automations to build, in order of impact, and for each one describe: the trigger, the action, and the specific friction it removes.”
This prevents the common mistake of automating for its own sake. The AI’s discrimination between data-gathering (automate) and thinking (don’t automate) is one of the most useful outputs a productivity assistant can provide.
Prompt 4: Weekly Review — Goal State Analysis
Use this every week once your SSoT and task log are current.
“Here is my goal status this week: [paste your SSoT summary — goals, current progress, status labels]. Here are my completed tasks this week, tagged by goal: [paste task completion log]. Here is my time allocation by goal for the week: [paste from time tracker — even rough estimates work]. Do three things: (1) Tell me which goals are on track, at risk, or stalled, with a one-sentence reason for each. (2) Identify any pattern across goals — are there consistent blockers, time drains, or behaviors showing up repeatedly? (3) Recommend the three most important tasks for next week that move my highest-priority goal forward. If the data is insufficient for any of these, tell me what’s missing rather than guessing.”
The instruction at the end — “tell me what’s missing rather than guessing” — is important. Without it, AI assistants tend to fill data gaps with plausible-sounding analysis that isn’t grounded in your actual situation.
Prompt 5: Quarterly Reset — Redesign My Connected Stack
Use this at the end of each quarter when goals change.
“I’m about to reset for a new quarter. My outgoing goals were: [list Q-outgoing goals and their final status — met / partially met / abandoned]. My incoming goals are: [list Q-incoming goals with success criteria]. Here are the tools in my current stack: [list tools]. Please help me: (1) identify which connections in my stack I should keep for the new quarter, (2) flag any connection that served last quarter’s goals but doesn’t serve this quarter’s, (3) identify any new connection I should build for the incoming goals, and (4) suggest whether any goal in my new list is likely to be under-resourced given my tool setup and the capacity I’ve described. Give me a concrete connection checklist for the first week of the new quarter.”
This prompt treats the quarterly reset as an architecture review, not just a goal-list refresh. It surfaces the tool-level changes that new goals require — changes that most people skip because they’re not visible until the goal stalls.
Pick the prompt that matches your most immediate need and run it with real data from your current system today.
Tags: AI prompts tool integration, Claude goal prompts, weekly review prompt, automation design AI, productivity stack prompts
Frequently Asked Questions
-
Do these prompts work with any AI assistant, or only Claude?
They're designed for Claude but work with any capable AI assistant — ChatGPT, Gemini, or others. The key is providing the structured data that each prompt calls for. The AI's output quality is a function of the input quality: the more specific and structured your goal data and tool inventory, the more specific and useful the AI's analysis.
-
How do I know if my data is detailed enough for these prompts?
A useful test: would a smart colleague who didn't know your situation understand what you're trying to accomplish from the context you've included? If not, add more. The most common omission is goal success criteria — what 'done' or 'on track' looks like for each goal. Without that, the AI can analyze patterns but can't tell you whether the patterns are good or bad.