Problem Framing with OCAREO

Why this exercise?

Most people talk to AI like they talk to a colleague — vague, incomplete, hoping the other side fills in the gaps. AI can't fill in the gaps. It doesn't know your context, your constraints, or what "good" looks like to you. The result? Generic answers that feel like they were written for anyone.

OCAREO is a 6-component framework that fixes this. It turns a loose idea into a precise specification — and the difference in output quality is night and day.

What is OCAREO?

Six components. Each one eliminates a type of ambiguity:

Letter	Component	What it does
O	Objective	Defines WHO the AI should be + WHAT it must achieve
C	Context	Paints the full scenario — industry, constraints, audience
A	Actions	Lists the exact steps the AI must follow
R	Rules	Sets what to NEVER do and ALWAYS do
E	Examples	Shows a concrete input-to-output transformation
O	Output	Defines the exact format of the final response

Think of it as a briefing document for your AI. The more precise the briefing, the better the result. No AI can compensate for a vague instruction.

Step 1 — Install the skill

Download the skill file above and install it in ToqanClaw:

1. Download problem-framing.zip using the link above
2. Open ToqanClaw and go to the Skills tab
3. Drag the .zip file into the Skills tab — it installs automatically
4. The skill is ready to use

Step 2 — Frame a real problem

Open a new conversation in ToqanClaw (click the +) and try this:

I want to explore a problem using OCAREO.

Here's my situation: I need to improve how my team
handles weekly status updates. Currently everyone sends
a different format, some people forget, and by the time
I compile everything for leadership it's Friday afternoon
and half the data is stale.

Help me frame this properly.

The skill will:

1. Analyze your request and identify gaps
2. Ask clarifying questions (team size? tools? what leadership wants?)
3. Build each OCAREO component with you
4. Deliver a complete, ready-to-use prompt

Step 3 — Compare before and after

Here's the power of OCAREO. Run both prompts below in ToqanClaw and compare what you get.

Without OCAREO (looks reasonable — breaks everywhere)

You're a helpful assistant. Write a quarterly risk summary
for our leadership team. Cover the main compliance risks,
keep it professional and board-ready. Suggest next steps
for anything critical. We're a mid-size fintech.

This prompt looks thoughtful. It has a role, a tone, a format request, and context. Most people would think this is good enough.

What the AI actually produces:

• Invents 5 "compliance risks" it found plausible (none are yours)
• Cites regulations from the wrong jurisdiction (US instead of EU)
• Suggests "consult legal counsel" for everything (no specificity)
• Writes 3 pages instead of the 1-page summary leadership expects
• No severity levels, no owners, no timelines

The AI filled every gap with confident-sounding guesses. If you skim it, it looks professional. If someone who knows your business reads it, it's fiction.

With OCAREO (clear, specific, usable)

OBJECTIVE: You are a compliance risk analyst who writes
concise board-ready summaries for fintech leadership.

CONTEXT: 200-person fintech, EU-regulated (PSD2/GDPR),
operating in NL, DE, PT. Q2 known issues: two GDPR
data subject requests exceeded 30-day deadline; one
PSD2 audit finding on transaction monitoring gaps.
Risk appetite: conservative. Audience: CFO + General
Counsel. They know the business; don't over-explain.

ACTIONS:
1) Inventory all known Q2 compliance events
2) Classify by severity (Critical / High / Medium)
3) Assign an owner and a proposed mitigation per item
4) Draft a 1-page board narrative connecting the dots

RULES:
NEVER invent compliance events — only use what I provided.
NEVER recommend legal strategy (that's counsel's job).
ALWAYS flag items marked [UNKNOWN] instead of guessing.
ALWAYS include severity + owner + timeline per risk.
Keep it under 400 words.

EXAMPLES:
Input: "GDPR SAR exceeded 30-day deadline (April)"
Output: "Severity: High | Owner: DPO (van Dijk) |
Mitigation: implement automated SAR tracker by Q3.
Status: remediation in progress."

OUTPUT: 1-page memo. Sections: Executive summary (3 lines),
Risk inventory (table), Recommended actions (numbered),
Next review date.

What the AI produces: Real risks from your Q2 data. Correct jurisdiction. Severity levels with owners. Under 400 words. Board-ready.

The difference is not a smarter AI. It's a clearer briefing.

Don't want to type? Talk it out.

If writing a full OCAREO feels heavy, you don't have to start from a blank page. Record a voice note or have a quick conversation about your problem, then use the Talk to Spec skill to extract the OCAREO structure from the transcript.

How it works:

1. Record a voice note (solo) or a short conversation (pairs/groups) about your problem — 5-15 minutes
2. Get the transcript (your phone's voice recorder, Otter, Fireflies, or any transcription tool)
3. Paste it into ToqanClaw and say "talk to spec"
4. Review the draft — correct anything marked [INFERRED], fill in anything marked [MISSING]
5. You now have a structured OCAREO spec from a conversation

You speak. The skill writes the spec. You review and correct. That's it.

Step 4 — Use it on YOUR real problem

Now pick something real from your work. A problem you've been thinking about. Open a new conversation and say:

I want to explore a problem using OCAREO.

[Describe your real situation here — be as specific
as you can. The more context you give, the better
the framework will help you.]

Let the skill guide you through the six components. When it's done, you'll have a prompt that would have taken you 30 minutes to write from scratch — and it will produce results you didn't think AI could deliver.

OptionalGo deeper▸

Once you're comfortable with OCAREO, try these advanced moves:

• Chain OCAREO prompts — use the output of one as the context for the next
• Template your recurring tasks — build OCAREO templates for things you do weekly
• Teach your team — share the framework so everyone writes better prompts
• Combine with my_persona — your persona provides the Context automatically

What just happened

You learned the single most impactful skill in AI: how to communicate precisely with a model. Every other exercise in this program produces better results when you apply OCAREO to it. Morning briefings, presentations, data analysis — they all improve when the instruction is structured.

OCAREO is the Portuguese methodology OCANES adapted for English: Objective, Context, Actions, Rules, Examples, Output. Same framework, same power.