Finding Your Use Cases: The Brain Dump and Reverse Prompt

Most people approach automation backwards.

They look at OpenClaw, think "what can this do?", browse the feature list, and try to map it to their life. This produces generic, low-leverage automations. Daily briefings about topics you do not actually care about. Automated tweets that do not sound like you. Cron jobs that run but produce nothing you act on.

The right approach inverts it.

The Brain Dump

A brain dump is a structured data transfer from your head to your agent's context. It is not a vague overview — it is a detailed, specific account of your goals, your daily workflows, your recurring frustrations, and the outputs you produce.

Here is what to include:

Professional context:

• What do you do and for whom?
• What decisions do you make regularly?
• What information do you need to make those decisions?
• What reports or outputs do you produce on a cadence?

Personal goals:

• What are you trying to build in the next 12 months?
• What content are you creating or want to create?
• What knowledge domains do you monitor?

Daily workflow:

• Walk through your actual day — what do you do at 8 AM, 12 PM, 5 PM?
• What tasks do you do every day that follow the same pattern?
• What do you procrastinate on because it is tedious but important?

Pain points:

• What takes too long?
• What falls through the cracks because you forget?
• What do you wish "just happened" without you having to initiate it?

Write all of this. Do not filter for what sounds automatable. Dump everything. The agent filters — that is the next step.

The Reverse Prompt

Once you have the brain dump written (500–1500 words is typical), paste it into OpenClaw and use this prompt:

"Based on everything you now know about me, my goals, and my daily workflows, what are the 10 most impactful automations you could implement for me right now? For each one: name, description, how it works, what pattern it follows (cron/trigger/on-demand), which channel it delivers to, and what the success metric is."

OpenClaw will generate a prioritized automation roadmap tailored to your actual context. Not generic examples. Your workflows.

The Three Tiers of Use Cases

The automations you will discover fall into three tiers. Your roadmap should have all three.

Beginner: Scheduled Intelligence Delivery

These are the foundation. They prove the system works and deliver immediate value.

• Daily briefs (covered in Lesson 133)
• Market price alerts
• Competitor monitoring summaries
• Weekly digest newsletters
• Calendar and task reminders with context

Pattern: gather → synthesize → deliver on schedule. No downstream triggers. Single output per run.

Intermediate: Chained Workflows

These are where compounding begins. One automation triggers the next.

• Brief → tweet draft → Discord review queue
• Market alert → trade hypothesis → Notion research page
• Weekly digest → LinkedIn thread → newsletter teaser
• Code commit → automated review → Slack notification

Pattern: each output is also a trigger for the next step. The agent hands off between stages.

Advanced: Intelligence-to-Action Loops

These are the highest-leverage automations — ones where the agent not only delivers intelligence but acts on it.

• Monitor news for pain points in your market → automatically generate a product concept → create a Notion spec → draft a landing page prototype
• Track competitor pricing → detect an anomaly → generate a strategic memo → schedule a meeting
• Monitor your own content performance → identify the highest-performing themes → auto-generate follow-up content in the same vein

Pattern: intelligence loop with active execution. The agent makes judgment calls and takes action, not just delivery.

These require more trust calibration — you want guardrails and review gates before the agent publishes or executes anything consequential. Build them after the simpler tiers are running reliably.

One Feature at a Time

The biggest mistake after running the brain dump is trying to build all 10 automations at once.

Do not.

The pattern:

1. Pick the highest-value automation from your roadmap
2. Build it in isolation: "I want to build [automation X]. Here is the context: [minimal relevant context]. Build only this."
3. Verify it runs correctly on the first scheduled execution
4. Then open a new conversation: "I have [automation X] running. Now I want to add [automation Y]."

The isolation discipline is non-negotiable. It is the difference between a system that compounds and a system that breaks silently.

My Automation Roadmap (Anonymized)

When I ran this exercise two years ago, my brain dump surfaced these as the top 10:

1. Daily macro brief → implemented week 1
2. Content performance monitoring → implemented week 2
3. Competitor move tracker → implemented week 3
4. Trade alert system → implemented month 2
5. Blog Autopilot → implemented month 2
6. Smart Engage (X + Instagram) → implemented month 3
7. Multi-agent advisory system (multi-perspective decision review) → implemented month 4
8. Platform health monitor → implemented month 4
9. Weekly newsletter pipeline → implemented month 5
10. Devlog pipeline (engineering content pipeline) → implemented month 6

Twenty-four weeks from brain dump to 10 running automations. None of them required me to rebuild from scratch — each one built on patterns established by the last.

That is the compound architecture in practice.

Lesson 134 Drill

Run the brain dump today:

1. Write 500–1500 words on your professional context, goals, daily workflow, and pain points
2. Paste it into OpenClaw via Telegram or Discord with the reverse prompt
3. Review the 10 proposed automations — are they specific to your actual workflow?
4. Rank them by impact and feasibility
5. Pick the top one and add it to your implementation queue

The output of this drill is your OpenClaw implementation roadmap. Keep it — you will refer back to it for the next several months.