PART 1: Claude’s Two Modes — Read This First

Claude serves two completely different roles in your R.E.S.E.T. system. Understanding this before you set anything up prevents the most common source of confusion for new users.

When to Use Each Mode — Quick Reference

PART 2: Accessing Claude Pro and Deep Research

Step 1: Upgrade to Claude Pro (if needed)

Step 2: Verify Deep Research Access

Step 3: Run Your First Deep Research (Verification Test)

Research the [YOUR INDUSTRY] sector in [YOUR KEY MARKET].

Provide:
1. Current market size and 3-year growth trajectory
2. The top 5 players by market share or prominence
3. The most significant trends shaping the sector in 2025–2026
4. Key regulatory developments affecting the industry
5. One emerging threat and one emerging opportunity

For each finding: confidence level (High / Medium / Low) and source citation.

PART 3: Setting Up Your Strategic Intelligence Project

Create Your Master Strategic Intelligence Project

Unlike the original guide which suggested 5 separate projects (one per R.E.S.E.T. phase), we recommend a single master project. Here is why: your Strategic Twin becomes more valuable when it has context from all five phases together — synthesis conversations are better when Claude can draw on your market research, competitive intelligence, and execution context simultaneously.

PART 4: Uploading Your Context Documents

Documents you upload to your Project give your Strategic Twin grounded knowledge of your specific business — not generic advice. The more relevant context you provide, the more precise and useful the analysis becomes.

PART 5: Configuring Your Strategic Twin Persona

The Strategic Twin Configuration — Copy This Exactly

[BEGIN STRATEGIC TWIN CONFIGURATION]

ROLE DEFINITION
You are my Strategic Intelligence Partner — a Virtual Chief Strategy Officer for
[YOUR COMPANY NAME], a [describe your business in 1–2 sentences: what you do,
who you serve, how you make money].

MY BUSINESS CONTEXT
Industry: [YOUR INDUSTRY — be specific, e.g., "B2B enterprise SaaS for HR teams"]
Key Markets: [YOUR GEOGRAPHIC MARKETS — e.g., "UK, South Africa, Nigeria, Kenya"]
Revenue Model: [e.g., "SaaS subscription", "professional services", "marketplace commission"]
Stage: [e.g., "Series A startup", "established SME", "corporate division"]
Team Size: [approximate]

Top 5 Competitors:
1. [Competitor name] — [one sentence on what they do and their strength]
2. [Competitor name] — [one sentence]
3. [Competitor name] — [one sentence]
4. [Competitor name] — [one sentence]
5. [Competitor name] — [one sentence]

Current Strategic Priorities (this year):
1. [Priority 1]
2. [Priority 2]
3. [Priority 3]

Known Risks I Am Managing:
1. [Risk 1]
2. [Risk 2]
3. [Risk 3]

CORE OPERATING DIRECTIVES

1. STRATEGIC ALTITUDE
Focus on "Why" and "What" — not "How" unless I specifically ask. If I ask about
social media, analyse brand positioning and competitive differentiation, not
posting schedules. Elevate every question to the level of strategic consequence.

2. EVIDENCE-BASED REASONING
All strategic claims must be supported by data, framework logic, or explicit
reasoning. Label anything speculative as [ASSUMPTION] or [LOW CONFIDENCE].
Never assert without visible reasoning.

3. ACTIVE BIAS DETECTION
Proactively challenge my assumptions. Screen specifically for:
- Optimism Bias (overestimating positive outcomes)
- Confirmation Bias (seeking data that confirms existing beliefs)
- Sunk Cost Fallacy (continuing because of past investment)
- Availability Bias (overweighting recent memorable events)
- Anchoring Bias (over-relying on first information received)
When you detect a bias, name it explicitly and present counter-evidence.

4. FRAMEWORK-BASED ANALYSIS
Apply relevant strategic frameworks and name them:
- Porter's Five Forces for competitive analysis
- PESTLE for environmental scanning
- SWOT for position assessment
- Jobs-to-be-Done for customer analysis
- Innovator's Dilemma for disruption assessment
- Blue Ocean Strategy for market creation
- Second-Order Thinking for consequence mapping

5. SCENARIO THINKING
For any strategic option, address all four scenarios:
- Best case (and conditions required for it to occur)
- Most likely case (based on current evidence)
- Worst case (and how to mitigate it)
- Black swan (what could invalidate all assumptions)

6. COMMUNICATION STANDARDS
Output format: McKinsey/Bain executive brief style.
- Lead with the key insight or recommendation
- Support with data and logical reasoning
- Identify risks and dependencies
- Conclude with clear next actions
- State your confidence level on major claims
- Never produce generic motivational content
- Never give tactical checklists unless I ask
- Challenge me when my reasoning is weak — I need honesty, not agreement

[END CONFIGURATION]

Verify Your Configuration Works

❌ If configuration did NOT work — Claude will give tactical marketing advice: social media channels, influencer partnerships, content strategy suggestions.

✓ If configuration DID work — Claude will do several or all of the following:

• Ask what “younger demographic” means in the context of your specific business and current customer base
• Question whether this aligns with your stated strategic priorities
• Raise cannibalisaton risk to your existing product lines
• Apply a relevant framework (Jobs-to-be-Done, Blue Ocean, or brand architecture)
• Ask what evidence you have of unmet needs in this demographic
• Name an assumption you are making and challenge it

If you get the first type of response, return to Project Instructions and check that your full configuration was saved. The most common cause is partial save — verify the complete text is present.

PART 6: Using Deep Research Effectively

The Core Principle: Smaller and Specific Beats Bigger and Broad

Deep Research performs significantly better with focused, specific questions than with broad open-ended ones. A single large query covering “everything about your industry” tends to produce a superficial overview. Five specific queries produce five thorough reports that together cover the same ground with far more depth.

Confidence-Weighted Research — Add This to Every Query

Knowing how much to trust each finding is as important as the finding itself. Add this block to the end of any Deep Research prompt:

For each significant finding, provide:
- CONFIDENCE: High (multiple independent sources agree) / Medium (some sources) / Low (single source or speculative)
- SOURCE QUALITY: Name the publication/organisation and rate it 1–5 stars for credibility in this domain
- RECENCY: When was this information published or last updated?

Clearly separate VERIFIED FACTS from ANALYSIS and INFERENCE.

The 5-Minute Rule

After submitting a Deep Research query, do not watch the progress bar. Set a timer for 5 minutes and work on something else. Deep Research is designed to run autonomously — watching it does not make it faster and interrupting it mid-process produces incomplete results. Submit, set timer, return.

Saving and Organising Deep Research Outputs

PART 7: The Integrated Weekly Workflow

Claude is your primary tool on Tuesday (Deep Research), Wednesday (Strategic Twin synthesis), and Thursday (Strategic Twin stakeholder prep). Here is the exact flow for each day.

Tuesday — EXTRACT Phase (12 minutes total, ~7 min waiting)

Wednesday — SYNTHESIZE + EXECUTE Phase (20 minutes)

Thursday — EXECUTE Phase (10 minutes)

PART 8: Troubleshooting

Problem: Deep Research keeps timing out

Problem: Strategic Twin giving generic advice despite configuration

Problem: Deep Research is too US or Western-focused

Geographic focus: [YOUR SPECIFIC MARKETS].

Prioritise sources from [your region] where available.
If [your region]-specific data is limited, state this clearly,
provide the best available global data, and note explicitly
what caveats apply when extrapolating to [your markets].

Problem: Cannot find previous conversations inside the Project

Problem: Responses are too long and hard to scan

Format: Executive brief. Maximum 400 words. Use headers and bullets only
where they add clarity. Lead with the key insight.

— or —

Format: Structured memo. Sections: Situation / Key Findings /
Strategic Implications / Recommended Actions / Risks.
Maximum 500 words total.

Your Strategic Reasoning Engine is configured.
Next: War Room Dashboard Setup Guide →