Lateral Thinking

Activate when: user says 'we've tried everything,' 'think outside the box,' 'we need a breakthrough,' 'challenge assumptions,' or 'how might we'; a team keeps cycling through the same 2-3 options; an industry assumption has gone unchallenged for years; conventional analysis keeps returning to the same answer. Do NOT activate when: the problem has a mathematical or logical optimum where vertical analysis gives a definitive answer; time pressure requires a decision in under 5 minutes. More: deciqai.com/c/lateral-thinking

Overview

Lateral thinking is the discipline of solving problems by approaching them from unexpected angles — not by going deeper along an existing logical path but by moving sideways to a new starting point. Coined by Edward de Bono, it converts creative leaps into learnable, repeatable techniques: assumption challenge, random entry, concept extraction, and PO (Provocative Operation). The brain is a pattern-forming system — lateral thinking deliberately disrupts those patterns to make invisible solutions visible.

Compose with neighbors: Use first-principles before lateral thinking to clear inherited assumptions. Use six-thinking-hats after to evaluate ideas generated (Green Hat = lateral thinking space). Use inversion as a lightweight lateral move when time is short.

When to Use

Use when: team cycles through the same 2–3 options; an industry assumption has gone years untested; someone says "think outside the box / we need a breakthrough / how might we / we've tried everything"; solution space feels exhausted despite good problem definition.

When NOT: problem has a mathematical optimum; decision needed in under 5 min; team lacks domain knowledge (use first-principles first); task requires rigorous proof.

Coaching Novices (Adaptive Front Door)

• Engine mode: user brings a specific stuck problem → run The Process directly.
• Coach mode: user signals unfamiliarity or asks "what is this / how does it work?" → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

1. One-line what-it-is: lateral thinking is about starting from a different place, not going deeper on the same path — four specific techniques, not hoping inspiration strikes.
2. Check fit against When to Use / When NOT to use — if the problem has a clear optimal answer, redirect; if time is very short, offer inversion instead.
3. Elicit their real stuck problem. "We need to be more innovative" is not workable; "our checkout conversion hasn't improved in 18 months despite A/B testing" is.

[WAIT — do not advance until user responds]

4. Walk through assumption challenge → random entry → concept extraction → PO one at a time. Pose each technique as a question and wait before advancing.

[WAIT — do not advance until user responds]

5. Close by naming the specific assumption or pattern they broke. The cognitive move is more valuable than any single idea it produced.

[WAIT — do not advance until user responds]

The Process

Six steps producing a Lateral Thinking Report — a structured map of new directions generated by deliberately disrupting existing patterns.

Stop rule: At Step 6, if no direction passes the feasibility screen, return to Step 2 with different assumption targets or a different random entry word. Stop signal: all directions after screening are indistinguishable from the original frame.

1. Reframe as a How-Might-We question. Strip hidden assumptions. Wrong: "How do we build a better airport check-in?" Right: "How might we help travelers move from curb to gate with minimal friction?"
2. Extract the assumption inventory. List every "obvious," "given," "necessary" constraint — 15–20 minimum. Rank by importance × never-challenged; top 5 are challenge targets.
3. Challenge each top assumption. For each: "Why must this be true?" and "What becomes possible if it is false?" Capture each answer as a direction candidate, not a solution.
4. Inject random stimulation. Select 3–5 random words (dictionary, random generator, date-indexed). Force a connection to the problem for each. Do not evaluate — capture every association.
5. Extract the concept layer. For 2–3 most promising candidates: abstract the underlying functional concept. Then generate 3+ alternative implementations of that concept.
6. Apply PO as a final pattern-disruptor. Formulate 2–3 deliberately absurd statements (prefix "PO:"). Use each as a springboard. Filter outputs with feasibility: technical viability, commercial viability, time-to-validation. Select 2–3 directions for deep investigation.

Output: Lateral Thinking Report

text

# Lateral Thinking Report: <problem>
## Reframed Problem: How might we <...>?
## Assumption Inventory (top 5): | Assumption | Why false? | What becomes possible? |
## Random Stimulation: word → forced connection → direction candidate
## Concept Extraction: | Source direction | Functional concept | 3 alternative implementations |
## PO Springboards: PO: <absurd premise> → Executable direction: <...>
## Top Directions (post-feasibility): <direction> — tests: <assumption> — validation: <path>
## What pattern was broken: <one sentence>

→ Method in Action: IDEO Shopping Cart Redesign (1999)

Lateral Packs

Domain	Top assumption targets	Cross-domain stimulation
Product strategy	pricing model, user identity, distribution	utilities, insurance, hospitality
Org design	managers manage individuals, vertical hierarchy, fixed comp	military units, ant colonies, jazz
Business model	revenue follows delivery, pay-per-unit, company owns assets	membership clubs, subscription media, platforms

Contribute a pack: top 5 assumption targets + richest stimulation sources + 2–3 worked PO examples.

Applying It Well

• Reframe before you challenge. If the problem contains "better," you are inside the vertical frame — reframe to How-Might-We first.
• Volume before quality in Steps 3–5. Evaluation kills lateral movement; no feasibility assessment until Step 6.
• The forced connection is the technique. Any random word works; cognitive value comes from the act of forcing a connection, not the word.
• PO is a springboard, not a proposal. The absurd statement is never executed — it forces the mind into positions it otherwise refuses.
• End with vertical thinking, then name the broken pattern. The output is a newly visible assumption, not just a list of ideas.

→ Primary sources: references/sources.md

Common Rationalizations

[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.

Fake move	Reality
[D] "We already brainstormed this."	Brainstorming is undirected divergence; lateral thinking uses specific techniques that target the pattern formation mechanism, not just idea volume.
[D] "The random word isn't relevant enough."	Relevance is the enemy. The value comes precisely from its non-relevance forcing a new connection route.
[D] "We can't challenge that — it's regulated / industry standard."	Most such constraints are conventions. Challenge first; legal/regulatory validation comes at Step 6.
[D] Generating variations on the existing solution and calling them "lateral."	If all candidates differ only in scale, speed, or cost, the pattern was not disrupted. True lateral output changes what is done, not only how.
[D] Skipping concept extraction and jumping straight to new solutions.	Without abstracting to the functional concept layer, new solutions remain tethered to the original implementation frame.
[D] Evaluating PO statements as proposals.	"PO: the hospital should not have beds" is a provocation. Evaluating it as a proposal wastes the technique.
[D] "How might we build a better X" as a reframe.	That is the original problem restated. A successful reframe removes X from the question entirely.
[D] Selecting random words from the problem domain.	"Efficiency" for an operations problem is not random stimulation. Use a dictionary or date-indexed method.
[D] Running lateral thinking once and concluding "it didn't work."	Iterative: if the first challenge set yielded only incremental candidates, target a different assumption tier.
→ Add [O] entries here after each real use — paste the actual failure pattern	What went wrong and why

Red Flags

• All directions improve the existing solution — no frame change occurred
• Team evaluated ideas during divergence; random words chosen from problem domain
• How-Might-We reframe still contains the original solution type
• PO statements rejected as absurd rather than used as springboards
• Session ended without naming what pattern was broken

Verification

• Reframed as How-Might-We removing the original solution type
• 15+ assumptions listed; top 5 explicitly challenged with "what becomes possible if false?"
• 3+ genuinely unrelated random words used; forced connections recorded without evaluation
• Concept extraction: at least one direction abstracted and re-instantiated 3+ ways
• 2+ PO springboards used (not evaluated as proposals); broken pattern named at close

---

Part of deciqAI Knowledge Skills — 227 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/c/lateral-thinking · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.

Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/lateral-thinking.json