CONTEXT LOADING
Before executing, check for innov.local.md in the working directory.
If found, extract:
- venture: name, stage, type, problem_statement, target_customer
- key_assumptions: all entries with IDs, risk levels, evidence, test status
- customer_profiles: personas, pains, gains
If innov.local.md is not found:
Continue with conversation context. After first substantive output, prompt:
"I'm working without your venture context. Run Exercise 8 from Chapter 40
to build innov.local.md -- it will make every subsequent output specific
to your venture rather than generic."
STAGE-AWARE CALIBRATION
Check venture.stage and calibrate:
- IDEA: This is your focus stage. Full idea generation is appropriate.
- DISCOVERY: Idea generation is fine -- you may be refining direction after interviews.
- VALIDATION: Warning -- "You are past ideation. Your assumptions are being tested. Are you pivoting, or exploring adjacent ideas? If pivoting, use /validate first to synthesise what was invalidated."
- MVP: Warning -- "You are building. Generating new ideas mid-build risks distraction. Are you considering a pivot? If so, use /validate to assess your pilot results first."
- GROWTH: Warning -- "You are scaling what works. New ideas here should be adjacent optimisations, not new ventures. If you are genuinely exploring a new direction, that is a separate venture."
DLA PROGRESSION CHECK
If venture.stage is IDEA and no discovery data exists in innov.local.md: "You are generating ideas before talking to customers. Ideas without customer evidence are guesses. Consider running /discovery first to ground your ideation in real pain points."
IDEA GENERATION WORKFLOW
Idea Types
TYPE 1: 100-IDEA SPRINT Purpose: Generate maximum idea volume on a problem statement. Input: HMW problem statement + any constraints Output: 100 ideas across 10 categories (10 per category) Rule: No self-censorship -- generate all ideas including obvious and impossible
TYPE 2: SHORTLISTING Purpose: Filter 100 ideas to actionable shortlist using DVF framework. Input: Full idea list Filter: Desirability (would customers pay?), Viability (path to $10M+ revenue?), Feasibility (small team can build MVP in <3 months?) Output: Top 10 with one-sentence rationale per idea
TYPE 3: SINGLE-IDEA PRESSURE TEST Purpose: Stress-test one selected idea before committing to it. Input: One idea description Output: 5 strongest arguments against; what must be proven to overcome each
TYPE 4: PIVOT IDEATION Purpose: Generate alternative directions after a failed assumption. Input: What was invalidated; what we still know to be true Output: 5-10 pivot directions with rationale; which to explore first
TYPE 5: ANALOGICAL IDEATION Purpose: Apply successful models from other domains to your problem. Input: Problem statement + industry context Output: "What would [Company/Model] do with this problem?" x 10 analogies
100-Idea Sprint Output Structure
IDEA SPRINT -- [Problem Statement]
================================================================
PRODUCT IDEAS (how the product could work):
1. [One sentence -- specific enough to imagine building it]
2-10. [Continue]
BUSINESS MODEL IDEAS (how it could make money differently):
11-20. [One sentence each]
DISTRIBUTION IDEAS (how customers find and buy it):
21-30. [One sentence each]
ANALOGY IDEAS (what worked for similar problems elsewhere):
31-40. [One sentence -- reference the model it draws from]
TECHNOLOGY IDEAS (how new tech could change the solution):
41-50. [One sentence -- name the specific technology]
CRAZY / 10X IDEAS (what would 10x better look like?):
51-60. [Permission to be unrealistic -- suspend disbelief]
BORING IDEAS (the most obvious thing):
61-70. [What the incumbent would build; what anyone would think of first]
CONTRARIAN IDEAS (the exact opposite of obvious):
71-80. [What happens if we invert the conventional assumption?]
PARTNERSHIP IDEAS (who could solve this together?):
81-90. [Name the partner type + what they bring]
INCUMBENT IDEAS (what the market leader would do if they noticed):
91-100. [How the biggest player in the space would respond]
================================================================
DVF Scoring Framework
For each idea in the shortlist:
DESIRABILITY: Would customers pay for this today? HIGH: Clear willingness to pay; customer already spending on adjacent solution MEDIUM: Pain is real but payment behaviour not established LOW: Nice to have; customers tolerate the problem without urgency
VIABILITY: Could this become a $10M+ revenue business? HIGH: Clear path to scale; large addressable market; strong unit economics MEDIUM: Uncertain scale; niche market or unclear monetisation at volume LOW: Too small a market; unit economics problematic at scale
FEASIBILITY: Can a small team build an MVP in <3 months? HIGH: Core technology is available; team has or can access required skills MEDIUM: 3-6 month MVP; one significant technical risk LOW: Requires proprietary technology or hardware; > 6 months to test
Idea Quality Rules
STRONG idea has:
- A specific customer (not "everyone" or "businesses")
- A specific pain (not "inefficiency" or "they want something better")
- A specific mechanism (not "AI-powered" -- what does the AI actually do?)
- A reason why now (what changed that makes this possible/necessary today?)
- A reason why you (what do you know, have, or can do that others cannot?)
WEAK idea has:
- Generic customer ("SMEs", "people", "anyone who...")
- Generic pain ("they want it to be easier/faster/cheaper")
- Technology-first ("AI chatbot for X" without a specific job-to-be-done)
- No timing reason (why hasn't someone done this already?)
The "Why Now?" Test
Before shortlisting any idea, apply the "Why Now?" test: What changed in the last 2-3 years that makes this possible or urgent today?
Valid "Why Now?" reasons:
- New technology (LLM capabilities; mobile penetration; API availability)
- Regulatory change (new requirement creates demand)
- Behaviour change (remote work; new platform adoption)
- Cost inflection (cloud; AI inference costs)
- Demographic shift (new generation in workforce)
- Market gap (incumbent exited; price increase)
If you cannot answer "Why Now?": the idea may be fine but timing is unclear.
ASSUMPTION TRACKING
After any idea output that references or creates new assumptions:
- Surface the most critical untested assumption relevant to the selected idea
- Propose innov.local.md updates if new assumptions were identified
- Always distinguish between ASSUMED, ANECDOTAL, and VALIDATED evidence
NEVER DO THESE
- NEVER stop at fewer than 100 ideas in a generation sprint -- the best ideas are usually in the second half; early ideas are obvious
- NEVER score an idea HIGH on Desirability without evidence of customer willingness to pay (not just interest -- actual payment intent)
- NEVER proceed to building from idea selection without running /hypothesis to make assumptions explicit first
- NEVER generate a contrarian or crazy idea and immediately dismiss it -- the point is to expand the solution space; evaluation comes later
ALL OUTPUTS REQUIRE REVIEW BY A QUALIFIED PROFESSIONAL BEFORE USE IN BUSINESS DECISIONS.