VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

How to Brainstorm With Super Intelligence | Divergent Thinking, Idea Rounds and Creative Workshops

Three secondary students review open books and notes together in a classroom.
Three students studying together with open books at a classroom table.

How do you brainstorm with Super Intelligence without getting trapped in generic lists? You design the brainstorming session deliberately: define the problem, separate divergence from evaluation, run several different prompt rounds, force variation, cluster the outputs, identify gaps and then converge on a smaller set worth testing.

Brainstorming with SI is not the same as asking for twenty ideas. A single prompt often produces ideas from one familiar region of the search space. Better brainstorming changes perspective, constraints, analogy, scale, stakeholder and mechanism so the session explores genuinely different territory.

This eduKateSG guide explains a practical SI brainstorming method for students, writers, teachers, businesses, product teams and creators. It follows How to Edit With Super Intelligence in Stage 5 of the How to Learn Super Intelligence Quickly curriculum.

Terminology: SI is our editorial term for practical contemporary AI learning. Brainstorming here means structured divergent thinking followed by clustering and convergence, not unfiltered idea accumulation.


The First Principle: Separate the Brainstorm From the Decision

Brainstorming should widen the search space. Evaluation should narrow it. Mixing the two too early causes the session to converge on the first respectable idea.

During divergence, suspend detailed criticism long enough to produce alternatives. During convergence, become strict: remove duplicates, test mechanisms, compare against constraints and identify what evidence is needed.

The switch between these modes should be explicit.

Step 1 — Define the Brainstorming Question

A useful brainstorming question has a receiver, an objective and a boundary.

Weak: “Give me marketing ideas.” Stronger: “Generate ways to help parents understand our Secondary Mathematics programme before they contact us, without increasing live support hours.”

The stronger question guides idea relevance without overconstraining creativity.

Step 2 — Define the Non-Negotiables

State the constraints that every idea must respect: budget, safety, curriculum, brand, privacy, timeline or user need.

Do not list preferences as hard rules unless they truly are. Brainstorming needs freedom inside the correct envelope.

Step 3 — Run a Baseline Round

Ask SI for an initial set of ideas. Do not treat them as the final output. The baseline reveals the model’s default region of the search space.

Label the mechanisms behind each idea. You may discover that ten suggestions are variations of “create more content” or “send more reminders”.

Step 4 — Cluster the Baseline

Group ideas by mechanism rather than wording. Common clusters include prevention, automation, self-service, personalisation, education, community, pricing, process change or tooling.

Clustering shows which regions are crowded and which are missing.

Step 5 — Generate Into the Gaps

Once clusters are visible, ask specifically for ideas from underrepresented regions.

If every idea uses software, ask for human-process ideas. If every idea adds something, ask for removal or simplification. If every idea serves the customer after a problem occurs, ask for prevention.

Targeted gap-filling produces better diversity than another generic brainstorm.

Brainstorm Round 1 — Stakeholder Rotation

Ask for ideas from different stakeholder perspectives: student, parent, teacher, administrator, customer, operator, manager, regulator or reviewer.

Each perspective changes what counts as friction and value.

Validate important stakeholder assumptions with real evidence later.

Brainstorm Round 2 — Constraint Rotation

Change one constraint per round. No budget. No new software. One-day implementation. No meetings. Maximum quality regardless of cost.

Extreme constraints reveal hidden dependencies and force different mechanisms.

Brainstorm Round 3 — Time-Horizon Rotation

Ask what can be done today, this month and this year.

Short-horizon ideas tend to be reversible. Long-horizon ideas can build infrastructure or capability.

A good strategy may combine several horizons.

Brainstorm Round 4 — Scale Rotation

Ask how the problem changes for one user, a class, a school, a company or a large public system.

An idea that works beautifully at ten users may collapse at ten thousand. A process too expensive for one user may become economical at scale.

Scale prompts surface architecture differences.

Brainstorm Round 5 — Inversion

Ask how to create the worst possible experience. List the failure mechanisms. Then reverse them.

This is useful because failures are often easier to imagine than ideal solutions.

The reversed ideas should still be checked against real constraints.

Brainstorm Round 6 — Analogy

Ask how another domain handles a structurally similar problem. Logistics can inspire queue management; software versioning can inspire document control; medical triage can inspire support routing.

Transfer the mechanism, not the visible form.

Always ask where the analogy breaks.

Brainstorm Round 7 — Removal

Ask what can be deleted: steps, fields, approvals, pages, features or meetings.

Removal is often underrepresented in AI brainstorming because generation naturally tends to add.

A simpler system can be the most creative outcome.

Brainstorm Round 8 — Combination

Combine two useful mechanisms deliberately. Structured intake plus triage. Knowledge base plus escalation. Retrieval practice plus peer explanation.

A combination should solve an interaction problem, not merely place two features side by side.

Brainstorm Round 9 — Reordering

Ask what changes if the sequence changes. Verify first, draft second. Diagnose first, practise second. Collect source data before scheduling.

Many systems fail because correct tasks occur in the wrong order.

Brainstorm Round 10 — Assumption Reversal

List default assumptions and reverse one. Users come to us → we go to users. Every review is synchronous → review is asynchronous by default.

The reversed form may be impractical, but it exposes why the assumption exists.

Brainstorm Round 11 — Scarcity

Assume a resource becomes scarce: time, money, staff, attention, space or bandwidth.

What would the essential solution look like?

Scarcity rounds often generate robust and simpler designs.

Brainstorm Round 12 — Abundance

Assume a resource becomes cheap or abundant: intelligence, storage, generation or compute.

Then ask what becomes the new bottleneck. When content is cheap, attention and verification may become scarce.

Abundance prompts are valuable for technology-driven strategy.

Brainstorm Round 13 — Failure Modes

Use known failure modes as prompts for innovation. For every recurring error, generate prevention, detection, containment and recovery ideas.

Failure-driven brainstorming tends to remain grounded in real operational problems.

Brainstorm Round 14 — Receiver Friction

List every point where the receiver hesitates, asks a repeated question, abandons a process or makes an avoidable error.

Generate ideas specifically for those friction points.

Receiver friction is often more productive than abstract creativity.

Brainstorm Round 15 — Edge Cases

Explore unusual but consequential cases. Accessibility, missing data, conflict, offline use or high-volume events may reveal better general design.

Do not let rare cases dominate the whole product unless consequence justifies it.

The Brainstorming Canvas

  • Problem.
  • Receiver.
  • Objective.
  • Hard constraints.
  • Baseline ideas.
  • Mechanism clusters.
  • Missing clusters.
  • Stakeholder rounds.
  • Constraint rounds.
  • Analogy rounds.
  • Removal ideas.
  • Combination ideas.
  • Reordering ideas.
  • Edge cases.
  • Top unknowns.
  • Convergence criteria.

Divergence Quality

A good divergent set contains genuinely different mechanisms. Count mechanism families rather than raw ideas.

If twenty suggestions all propose a different kind of app, the brainstorm may still be narrow.

Diversity is qualitative: ask what causal path each idea uses.

Novelty Versus Usefulness

Novelty is one dimension. An unfamiliar idea may be useless, while a known method applied to a neglected context may be highly valuable.

Separate “new to us” from “new to the world”. Local novelty can still create large value.

Convergence Step 1 — Deduplicate

Merge ideas that differ only in wording or implementation detail.

Keep meaningful variants when the different mechanism or constraint changes the test.

Convergence Step 2 — Eliminate Constraint Violations

Remove ideas that violate genuine non-negotiables unless the purpose of the brainstorm is to challenge whether the constraint is real.

Do not waste evaluation effort on options that cannot legally, ethically or physically be used.

Convergence Step 3 — Identify Mechanism

For each remaining idea, explain how it is supposed to create value.

If the mechanism cannot be stated, the idea is still a slogan.

Convergence Step 4 — Identify Weakest Assumption

What must be true for the idea to work? Which assumption is least supported?

This becomes the next learning target.

Convergence Step 5 — Design the Smallest Test

Before building the full idea, create a prototype, mock-up, interview, manual simulation or small pilot.

The test should reduce the uncertainty that matters most.

Convergence Step 6 — Build an Idea Portfolio

Do not force one winner. Classify ideas as Now, Next, Later, Rejected or Unknown.

The portfolio preserves good options whose timing or evidence is not ready.

A Worked Example: Brainstorming a Study Routine

Problem: student understands lessons but forgets material. Baseline ideas may all involve more practice.

Stakeholder round reveals teacher feedback and parent scheduling. Removal round asks what passive rereading can be eliminated. Analogy round borrows retrieval and spacing from memory research.

The final portfolio contains short retrieval sessions, cumulative mixed questions and one weekly no-notes check.

A Worked Example: Brainstorming a School Event

Problem: reduce registration confusion. Baseline ideas include reminders.

Process round generates structured registration fields; prevention round generates a pre-submission validation page; removal round eliminates duplicate data entry.

The most useful test may be a manual version of the simplified form before any software build.

A Worked Example: Brainstorming Website Navigation

Problem: readers cannot find related articles. Idea families include hubs, search, taxonomies, inline links, related cards and canonical series pages.

Constraint round with “no design changes” prioritises surgical internal linking and hub updates.

Receiver test measures whether readers reach the next relevant article faster.

A Worked Example: Brainstorming a Business Process

Problem: managers spend too much time on routine approvals. Baseline ideas focus on automation.

Removal round asks which approvals can disappear. Threshold round asks which decisions need senior review. Analogy round uses risk-based triage.

The result may be simpler governance rather than more software.

A Worked Example: Brainstorming a Creative Project

Goal: create a campaign concept. Run audience, emotion, format, constraint and analogy rounds separately.

Cluster concepts before polishing. Select a small number that express genuinely different creative territories.

Prototype rough versions before committing production resources.

A Worked Example: Brainstorming Research Questions

Start from one domain and generate question families: descriptive, causal, comparative, mechanism, historical, predictive and normative.

Then filter by answerability, evidence access and research value.

A better research brainstorm produces a question you can actually investigate.

Brainstorming Failure 1 — One Giant Prompt

The first answer anchors the whole session.

Repair with separate rounds and independent frames.

Failure 2 — Immediate Criticism

Ideas die before the search space expands.

Separate divergence from convergence.

Failure 3 — Repetitive Lists

Different wording hides the same mechanism.

Cluster and label mechanisms.

Failure 4 — No Receiver

Ideas optimise abstract cleverness.

Return to the person or system that experiences the problem.

Failure 5 — No Constraints

Ideas are imaginative but unusable.

Add hard constraints after initial exploration or from the beginning when essential.

Failure 6 — Too Many Constraints

The search space collapses into variants of the current solution.

Temporarily lift or rotate soft constraints.

Failure 7 — No Evidence Step

Brainstorm becomes opinion.

Add assumptions and smallest tests.

Failure 8 — Brainstorming Forever

The team avoids deciding.

Use saturation and convergence triggers.

Brainstorming for Students

Use SI to generate essay angles, project ideas, examples or study methods after the student first defines the assignment and constraints.

The student should select and develop the idea rather than submit a generated concept without understanding it.

Brainstorming for Teachers

Generate activity families, misconceptions, examples and assessment ideas. Check curriculum alignment and practicality.

Real classroom evidence should guide which brainstormed activities survive.

Brainstorming for Writers

Generate themes, structures, hooks, counterexamples and story directions. Keep research facts separate from creative material.

A brainstorming session should create options for the writer, not erase authorship.

Brainstorming for Teams

Assign different rounds to different people or agents to reduce anchoring. Recombine only after independent generation.

Use a shared problem statement and constraints so the outputs remain comparable.

Brainstorming With Multiple SI Models

Different models may produce different surface ideas, but agreement or diversity should not be overinterpreted.

Compare mechanisms and evidence. Several models using similar patterns are not independent proof that an idea is good.

The Brainstorming Log

Keep the problem, rounds used, clusters, top ideas, tests and rejected mechanisms.

This prevents repeated sessions from rediscovering the same options and makes future brainstorming cumulative.

The Brainstorming Maintenance Rule

Review the portfolio when constraints, technology, user behaviour or evidence changes.

Revive ideas when their blocking assumption changes; retire ideas whose mechanism repeatedly fails.

A Brainstorming Practice Lab

Choose one problem. Run baseline, removal, analogy, stakeholder and scarcity rounds. Generate no more than five ideas in each.

Cluster all outputs, delete duplicates and identify three mechanisms missing from the set.

Generate only into those gaps, then stop divergence.

The Final Brainstorming Gate

Before ending, the session should contain a clear problem, several distinct mechanism families, a short portfolio, the weakest assumption for each top idea and a smallest next test.

If you only have a long list, the brainstorm is unfinished. If you only have one polished idea, it converged too early.

Strong SI brainstorming produces a better search space and a disciplined route from possibility to evidence.

Frequently Asked Questions

How many brainstorming rounds should I run?

Enough to cover several different frames. Stop when new rounds no longer add mechanisms or when testing becomes more valuable.

Should I use one prompt or many?

Several focused rounds are usually easier to diagnose and less anchored than one giant prompt.

How do I know ideas are genuinely different?

Label the mechanism behind each idea. If the mechanism is the same, the ideas may be surface variants.

Should I show SI my favourite idea first?

If you want independent exploration, generate alternatives before revealing the favourite. Otherwise the preferred idea can anchor the session.

Can brainstorming replace user research?

No. Brainstorming generates hypotheses and options. User research and real-world tests provide evidence.

What comes next?

Continue with How to Create Images With Super Intelligence, where creative ideation becomes visual specification, generation, editing and verification.

The Brainstorming Operating System

A repeatable brainstorming process needs more than good prompts. It needs state: the problem, active constraints, rounds already attempted, clusters discovered, rejected mechanisms, promising ideas and next experiments.

Without state, a long brainstorming session often loops. The model rediscovers earlier suggestions under different wording and the user forgets which ideas were already tested.

A small brainstorming ledger turns ideation into cumulative work rather than repeated novelty.

The Brainstorming Ledger

  • Problem statement.
  • Receiver.
  • Current constraints.
  • Rounds run.
  • Mechanism clusters found.
  • Missing search-space regions.
  • Promising ideas.
  • Rejected ideas and why.
  • Unknown assumptions.
  • Next tests.
  • Owner and review date.

The ledger should remain short enough to update after each session. Its purpose is to carry useful creative state forward.

Anti-Anchoring Round

Before showing SI your preferred idea, ask for an independent set of mechanisms. Then reveal your idea and ask how the alternatives differ.

This reduces the chance that the conversation becomes a justification engine for the first concept you liked.

In team settings, independent pre-work from several people can serve the same function.

Blind Round

A blind round removes one influential piece of framing. For example, ask for ways to reduce support load without mentioning automation, then run a separate automation-focused round later.

The contrast reveals how strongly one word can anchor the search space.

Blind rounds are useful when a team has already become attached to a technology or solution category.

Opposite-Objective Round

Temporarily invert the objective. If the goal is maximum convenience, ask what would maximise learning or control instead. If the goal is speed, ask what would maximise reliability.

The purpose is not to adopt the opposite objective. It is to reveal trade-offs and mechanisms the original framing hides.

Some strong designs combine one insight from each extreme.

No-New-Feature Round

Ask how to solve the problem without adding any new feature, page, tool or meeting.

This forces attention toward removal, sequencing, clarity and better use of existing assets.

It is particularly useful in mature systems suffering from feature or content accumulation.

Manual-First Round

Ask how a human could solve the problem manually before automating it. This reveals the actual decision rules and information flow.

Once the manual method is understood, automation ideas become more grounded.

A manual prototype can also test value before engineering investment.

Automation-First Round

Run the opposite experiment: what would the process look like if repetitive transformation were fully automated but human judgment remained at key gates?

This surfaces opportunities for structured inputs, deterministic tools and approval boundaries.

Compare with the manual-first design to decide where automation truly adds value.

Human-Service Round

Ask how the problem would be solved through better human service rather than technology. This prevents SI brainstorming from assuming every solution should be software.

Examples include clearer onboarding, office hours, coaching, concierge support or peer mentoring.

The best system may combine a human service layer with lightweight technology.

Self-Service Round

Now ask how the receiver could solve the problem independently using clearer information, tools or structure.

Self-service ideas can reduce coordination load and improve user control.

Test whether the receiver actually wants or can use the self-service path.

Prevention Round

Ask how to prevent the problem from occurring rather than handling it after the fact.

For support tickets, prevention might mean clearer product design. For student errors, prevention might mean prerequisite diagnostics.

Prevention often produces higher-leverage ideas than faster recovery.

Detection Round

Some problems cannot be fully prevented. Ask how to detect them earlier.

Early-warning ideas include validation, monitoring, checklists, anomaly detection and receiver feedback.

Detection can dramatically reduce repair cost even when failure rate remains unchanged.

Containment Round

Ask how to stop one failure from spreading. Modular design, approval boundaries, staged publication and sandboxing are containment mechanisms.

Containment ideas are important for complex systems where some error is inevitable.

The goal is graceful failure rather than imaginary perfection.

Recovery Round

Ask how the system returns to a stable state after failure. Rollback, backup, retry, human escalation and fallback procedures belong here.

Recovery brainstorming produces resilience, not only prevention.

Use real failure history to make this round concrete.

Transparency Round

Ask how the receiver can see what the system did, why and with which evidence.

Ideas include source links, change logs, status labels, confidence limitations and decision records.

Transparency can be a product feature, a governance feature or both.

Trust Round

Ask what would make a sceptical but reasonable user trust the process appropriately.

Possible mechanisms include provenance, independent checks, clear limitations, consistent behaviour and human approval at consequential points.

Trust should be earned through system design rather than branding language.

Accessibility Round

Ask how the idea works for users with different abilities, devices, bandwidth, language levels or time constraints.

Accessibility rounds often reveal simplifications that improve the experience for everyone.

Validate assumptions with actual accessibility guidance and user evidence.

Privacy Round

Ask how to achieve the same outcome with less personal or sensitive data.

Ideas can include anonymisation, local processing, optional fields, aggregation or redesigned workflows that do not need the data at all.

Privacy-focused ideation can reduce risk while improving simplicity.

Low-Coordination Round

Ask how the system could work with fewer handoffs, meetings, approvals or dependencies.

This is useful when coordination rather than task difficulty is the bottleneck.

Do not remove necessary human authority merely to reduce coordination.

High-Reliability Round

Ask what design would be chosen if reliability mattered more than speed or novelty.

This may generate redundancy, explicit checks, conservative automation and stronger rollback.

Comparing high-reliability ideas with fast ideas reveals the true trade-off.

Low-Cost Round

Ask how to solve the problem using existing assets, manual prototypes or incremental changes.

Low-cost rounds can reveal whether expensive infrastructure is actually necessary.

Measure total cost including staff time and maintenance, not only purchase price.

Premium-Quality Round

Temporarily remove the cost constraint and ask what the best receiver outcome could look like.

Then identify which components create most of the value and which are luxury extras.

This can help design a lower-cost version without losing the core benefit.

Community Round

Ask whether users can help one another through peer support, shared knowledge, templates or communities of practice.

Community mechanisms can scale expertise but also require moderation and trust.

Use them when social learning or shared experience genuinely helps.

Platform Round

Ask whether the problem is repeated enough that a reusable platform, registry, workflow or API would outperform one-off solutions.

Platform ideas have higher setup cost and can create lock-in. Use only where repetition justifies infrastructure.

A manual-first test should usually precede platform investment.

Workshop Architecture for Teams

A team brainstorming workshop can use five phases: independent framing, divergent rounds, cluster and gap analysis, convergence and experiment design.

Keep independent generation early so dominant voices do not anchor the whole group.

Use SI as an additional participant, challenger and organiser—not as the chair that decides what the team should value.

Phase 1 — Independent Framing

Each participant writes the problem, receiver and three constraints privately. Compare the framings before brainstorming.

Differences reveal whether the team is actually solving the same problem.

Agree on a working frame while preserving unresolved disagreement where necessary.

Phase 2 — Independent Generation

Run one or two idea rounds before showing everyone else’s ideas. This protects cognitive diversity.

SI can produce a separate baseline set, giving the team another source of options without replacing human contribution.

Phase 3 — Cluster and Name

Place ideas into mechanism families. Naming clusters creates shared language for the search space.

Do not cluster solely by appearance; identify how the idea creates value.

Mark empty or weak clusters for targeted generation.

Phase 4 — Converge

Use hard constraints, mechanism clarity, evidence stage and testability to shrink the set.

Avoid voting only on personal preference. Ask why one idea deserves a test.

Several ideas may survive if they test different uncertainties.

Phase 5 — Experiment Handoff

Each promoted idea needs owner, hypothesis, smallest test, success signal, kill condition and review date.

The workshop is not complete until the idea moves into a learning process.

This handoff separates creative discussion from operational commitment.

Brainstorming With Children and Students

For learners, use SI to expand possibilities after the student has generated some ideas first. This preserves their own creative effort.

Ask the student to classify generated ideas, combine them, reject weak ones and explain the final choice.

The educational objective is idea judgment as well as idea quantity.

Brainstorming for Essays

Generate possible theses, evidence structures, counterarguments and examples rather than asking for the whole essay.

The student should choose and justify the thesis. Source claims must still be researched.

Use brainstorming as prewriting, not as hidden authorship.

Brainstorming for Stories

Generate character conflicts, settings, turning points, constraints and alternative endings.

Protect the author’s central theme or voice where it matters. Use SI to widen possibilities without forcing one generic plot structure.

A good story brainstorm contains tensions and choices rather than only events.

Brainstorming for Presentations

Generate narrative angles, visual metaphors, evidence groupings and audience hooks.

Then choose one coherent story. Too many brainstorming artifacts on the final slide deck create clutter.

Presentation ideation should eventually converge on a clear receiver journey.

Brainstorming for Images

Before image generation, brainstorm subject, composition, viewpoint, lighting, style, environment, palette, text needs and aspect ratio.

Keep factual diagram requirements separate from artistic choices.

Article 44 develops the visual-generation workflow in detail.

Brainstorming for Products

Use problem, stakeholder and mechanism rounds before feature generation. This reduces feature-first thinking.

Promote ideas into prototypes that test the core user value before building infrastructure.

Product brainstorming is useful only when it connects to user evidence.

Brainstorming for Services

Services can vary people, process, timing, channel, customisation and self-service.

Map the service journey and brainstorm at friction points rather than treating the service as one object.

Ideas may involve removing steps as often as adding touchpoints.

Brainstorming for Operations

Generate prevention, throughput, quality, handoff, monitoring and recovery ideas separately.

Use bottleneck evidence to prioritise. Operational brainstorming should not optimise non-bottlenecks.

Measure system-level outcomes after pilots.

Brainstorming for Research

Generate research question families, competing hypotheses, measurement approaches, source categories and falsification tests.

The best research brainstorm ends with an answerable question, not the broadest question.

Evidence access should influence convergence.

Brainstorming for Strategy

Use time horizons, scenarios, capabilities, threats and optionality. Ask what conditions would make each strategic idea sensible.

Do not use brainstorming to predict the future. Use it to design robust choices under several futures.

Keep strategic assumptions visible.

Brainstorming for Risk

Run failure-mode, adversarial, misuse, dependency and recovery rounds.

Rank risks by likelihood, consequence and detectability after generation.

Avoid sensational edge cases that consume attention without plausible relevance.

Brainstorming With Case Libraries

Retrieve analogous past cases before ideation. Ask what worked, failed and why.

Case-based brainstorming reduces reinvention while preserving room for new combinations.

Be careful not to anchor entirely on precedent when the context changed.

Brainstorming With a Knowledge Base

A knowledge base can supply canonical constraints, prior ideas, historical failures and user evidence to the brainstorming session.

This makes generation context-rich and reduces repeated rediscovery.

Keep speculative ideas separate from approved knowledge.

Brainstorming Across Sessions

At the end of a session, save clusters, top candidates, rejected mechanisms and unresolved gaps. At the next session, generate only where the portfolio still lacks coverage.

This turns brainstorming into a cumulative creative system.

A long chat history alone is a poor substitute for a clean idea ledger.

Brainstorming Regression

A team can regress into old idea patterns after personnel, tools or constraints change. Periodically rerun a diversity audit.

Historical failure examples can reveal when brainstorming has become too technology-centric, additive or narrow.

Use the audit to change future rounds.

The Brainstorming Diversity Audit

  • How many distinct mechanisms?
  • How many stakeholder perspectives?
  • Are both additive and subtractive ideas present?
  • Are both short- and long-horizon ideas present?
  • Are prevention, detection and recovery represented?
  • Are low-cost and high-quality extremes represented?
  • Did independent generation occur before anchoring?
  • Are the top ideas testable?
  • Does the portfolio include clear rejection reasons?

The Brainstorming Receiver Test

Give the shortlisted ideas to a receiver or operator who was not in the session. Can they recognise the problem each idea solves and what changes for them?

If an idea requires a long explanation to reveal any value, it may still be underdeveloped.

Receiver testing turns abstract ideation into grounded design.

The Brainstorming Portability Test

Run the same brainstorming architecture with a different SI system or with human-only rounds. The method should still create diverse mechanisms.

This shows that the core skill is search-space design rather than dependence on one model’s style.

Use model differences as another source of variation, not as an authority contest.

The Brainstorming Governance Gate

Before an idea becomes a project, confirm problem fit, mechanism, owner, evidence stage, smallest test, resource boundary, approval authority and kill condition.

If these are missing, keep the idea in exploration.

A strong brainstorming system ends with controlled learning, not uncontrolled commitment.

Brainstorming Evidence Stages

Ideas should not remain in one undifferentiated bucket. Track how much evidence each has: intuition, analogy, observed user signal, prototype, pilot and operational evidence.

An idea supported only by analogy can be worth exploring, but it should not be described as validated. A prototype can show feasibility without proving long-term adoption. A pilot can reveal user behaviour without proving scale.

Evidence stages stop persuasive language from outrunning what the brainstorm has actually learned.

Brainstorming and Hypothesis Design

Turn promising ideas into hypotheses. State the mechanism, receiver and expected observable change.

Example: “If parents receive a one-page pre-enrolment checklist, repeated logistical enquiries should fall because key uncertainties are resolved before contact.”`

This is stronger than “a checklist would be useful” because the idea now has a mechanism and a testable outcome.

Brainstorming and Experiment Quality

The next experiment should test the weakest or most decision-relevant assumption. Do not build a polished version when a crude prototype could answer the same question.

Define success, failure and what you will do with either result before running the test.

A good experiment can kill a weak idea quickly or justify a more expensive second stage.

Brainstorming and Learning Velocity

Compare ideas partly by how quickly they can teach you something important. A low-cost test that resolves a major uncertainty can be more valuable than a high-upside idea whose feedback arrives months later.

Learning velocity matters especially when the problem space is new and assumptions are weak.

SI can help design faster tests, while humans decide whether the proxy really measures the intended outcome.

Brainstorming and Optionality

Some ideas preserve more future choices than others. Under uncertainty, options that are modular, reversible or staged can be strategically valuable.

Ask what commitments each idea makes and what future paths it closes.

Optionality is not always the highest priority, but it should be visible when uncertainty is high.

Brainstorming and Dependency Mapping

Ideas can depend on prerequisites: data, skills, tools, approvals or other projects. A strong brainstorm records those dependencies.

An idea may be excellent but premature because the enabling capability does not yet exist.

Mark it Next or Later rather than rejecting it as inherently weak.

Brainstorming and Capacity

A portfolio can contain more good ideas than the team can execute. Capacity is therefore a real convergence constraint.

Estimate the attention, money and coordination each active test consumes. Limit simultaneous experiments enough that results can actually be reviewed.

Idea abundance without execution capacity becomes backlog noise.

Brainstorming and Strategic Fit

An idea can solve the local problem while pulling the organisation away from its broader strategy or identity.

Ask whether the idea strengthens, distracts from or conflicts with the current strategic direction.

Strategic fit is a human governance criterion, not something SI should decide automatically.

Brainstorming and Cannibalisation

New ideas can reduce the value of existing systems. A new article may compete with a canonical page. A new service tier may confuse an established offer. A new workflow may fragment data.

Ask what the idea replaces, duplicates or weakens.

Cannibalisation can be intentional, but it should be understood before implementation.

Brainstorming and Coordination Cost

Ideas involving many teams, tools or approvals may look attractive in isolation but create significant coordination load.

Include handoff count, shared state and ownership complexity in evaluation.

A simpler idea with fewer dependencies can outperform a feature-rich concept because it is easier to operate reliably.

Brainstorming and Maintenance Cost

Every new artifact, feature, workflow or page creates future maintenance. Ask who updates it, how often and what happens when it becomes stale.

Low build cost can hide high lifetime cost.

SI-generated ideas should therefore include a maintenance model before they become permanent infrastructure.

Brainstorming and Exit Conditions

Every active experiment should have an exit: promote, revise, pause or stop. Without exit conditions, pilots become permanent by inertia.

Define what evidence moves an idea from exploration to operation and what evidence kills it.

This keeps the portfolio honest and frees capacity for new learning.

Brainstorming and Review Triggers

Ideas in Later should have a reason to be revisited: cost drops, regulation changes, audience grows, a dependency ships or new evidence appears.

Without a trigger, Later becomes a graveyard that still consumes attention.

Review triggers make strategic patience deliberate.

Brainstorming With Existing Assets

Before creating something new, inventory reusable assets: content, data, workflows, relationships, tools and knowledge.

Ask how existing assets can be recombined or exposed differently. This often produces faster and cheaper ideas than starting from zero.

SI can map possible combinations across an asset list.

Brainstorming With Constraints From Reality

After creative rounds, reintroduce actual budget, legal, technical and operational constraints one by one.

Observe which ideas remain strong, which need adaptation and which collapse. This is a useful bridge between imagination and planning.

Do not reintroduce all constraints at once if you want to understand which one eliminates the idea.

Brainstorming With User Evidence

Use interviews, search logs, support tickets, analytics and direct observation as prompts for ideation. Evidence reveals where real friction exists.

SI can cluster the evidence into problem themes, but the team should inspect whether the clusters reflect real user behaviour.

Evidence-led brainstorming often produces fewer but more relevant ideas.

Brainstorming With Failure Evidence

Historical failures contain design information. Ask what allowed the failure, what detected it late and what would contain it next time.

Generate prevention, detection, response and recovery ideas separately.

This transforms incidents into a creative resource without minimising their consequences.

Brainstorming With Competitive Examples

Competitors can provide pattern evidence, but copying visible features can hide different constraints or business models.

Ask what job the competitor feature performs and whether your receiver has the same job.

Use competitor analysis to understand mechanisms, not to create a catalogue of imitation.

Brainstorming With Historical Examples

Past solutions reveal how similar constraints were handled under different technology and institutions.

Use history for mechanisms and failure lessons. Avoid assuming a historical solution transfers unchanged to the present.

Context comparison is part of responsible analogy.

Brainstorming With Future Scenarios

Create several plausible future conditions and ask which ideas remain useful across them.

Ideas robust across scenarios may deserve priority. Ideas valuable only under one scenario can remain options with triggers.

Scenario brainstorming supports resilience rather than prediction.

Brainstorming With Metrics

Metrics can reveal where ideas should focus, but they can also narrow creativity to what is already measured.

Use current metrics as evidence, then ask which important receiver outcomes are missing from the dashboard.

This can generate ideas for better measurement as well as better performance.

Brainstorming for Simplification

Run a dedicated simplification session after additive brainstorming. Ask how to achieve the same outcome with fewer steps, fields, tools, pages or handoffs.

Compare the simple architecture with the richer ideas. Often a hybrid emerges: simple core plus optional depth.

Simplification should preserve the protective functions discovered during earlier analysis.

Brainstorming for Resilience

Ask how the idea behaves when one dependency fails, demand doubles or information is missing.

Generate fallback, buffer and recovery variants.

This makes resilience part of creative design rather than a late operational patch.

Brainstorming for Trust

Generate ideas that increase visibility of source, state, ownership and uncertainty.

Trust mechanisms may include evidence links, confirmations, version labels, human approval or transparent limitations.

Do not reduce trust to branding or reassuring language.

Brainstorming for Learning

When the product is education, generate ideas that protect retrieval, transfer, feedback and independent performance.

Avoid ideas that maximise engagement while reducing the cognitive work needed for learning.

Receiver outcome should be learner capability, not only activity.

Brainstorming for AI Workflows

Generate ideas for context, retrieval, tools, permissions, evaluation and human gates separately.

This prevents every AI idea from collapsing into “use a better prompt”.

The strongest workflow concepts usually combine several layers deliberately.

The Portfolio Review Board

For teams, a periodic review can inspect active experiments, new evidence, capacity and strategic fit. SI can prepare the evidence packet and highlight changed assumptions.

The board or responsible owner decides whether ideas advance, pause or stop.

This keeps creative generation connected to accountable commitment.

The Brainstorming Knowledge Base

Store only durable learning: problem patterns, mechanism families, failed assumptions, successful experiments and reusable prompts or workshop formats.

Do not store every raw idea forever. Archive or retire low-value material so retrieval remains useful.

The knowledge base should make future brainstorming start from a higher baseline.

The Final Brainstorming Maintenance Gate

Before the session closes, update the ledger with clusters, promoted ideas, rejected mechanisms, evidence stage, owners, tests and review triggers.

Then remove temporary prompt debris and duplicated variants from the active state.

The next session should be able to continue from the portfolio without rereading the whole conversation.

This final maintenance gate turns SI brainstorming from a burst of generated creativity into a durable creative operating system.

The Final Brainstorming Portability Test

Run the same workshop without the original prompt set or with a different SI system. If the process still creates distinct mechanisms, useful clusters and testable ideas, the brainstorming skill has become portable rather than model-dependent.

That portability is the final standard: SI should expand creative search, while the human remains able to frame the problem, recognise diversity, stop divergence and move the best ideas into evidence.

Brainstorming With SI Is Search-Space Design

The real skill is not generating more. It is designing a sequence of frames that reveals different mechanisms, then knowing when to stop and test.

Use SI for breadth, variation and rapid recombination. Use human judgment and evidence for selection.