2.4Brainstorm & structure your ideas
The essentials in 30 seconds
- Divergence calls for volume: twenty ideas in thirty seconds, three of them good — sort afterwards, never judge during.
- Imposed angles break the banal: zero budget, competitor, personas — first ideas are the average, later ones the value.
- The devil's advocate offsets complacency: attack the plan like a skeptic, then you decide with full knowledge.
Threaded mini-case — Amine and the solo ideation workshop in Algiers
Continuing the level: Amine, project manager in Algiers, must lead the next team meeting and improve client welcome. Alone before his blank page, he requests 20 ideas from the obvious to the unexpected, adds 5 zero-budget ideas and 5 ideas his competitor would dare, ranks everything in an impact/effort matrix with an argued top 3, then has his No. 1 idea demolished “like a skeptical CFO”. He decides himself, with full knowledge.
Teaching model: the quoted instructions illustrate divergence, convergence and stress-test on an ordinary topic.
1. A thinking partner
AI generates twenty ideas in thirty seconds (three of them good — that is the brainstorming principle), structures your scattered thoughts into a coherent plan, and plays devil's advocate on request. Like Amine before his blank page: the symmetric trap would be letting it think in your place — AI broadens YOUR thinking, it does not replace it.
2. Divergence, convergence, structuring
Divergence (quantity of ideas, imposed angles, personas), convergence (criteria, sorting matrices) and structuring (from scattered notes to action plan) form the solo ideation workshop trio. Ask for quantity, force variety (“5 zero-budget ideas, 5 digital ideas, 5 ideas a competitor would try”), then have them sorted by explicit criteria.
3. The stress-test before deciding
The on-demand devil's advocate (“attack this plan like a competitor, an unhappy client, a CFO”) offsets the model's natural complacency — which otherwise goes along with you. In a team: human ideas first (so as not to anchor the group), then AI to broaden and challenge. The final decision remains human.
🔎 Concrete example — 🔎 Worked example — The gem in the volume
The follow-up that escapes the banal (method taken from V2):
“20 ideas, from the obvious to the unexpected” (the 1st are average).
“+ 5 zero-budget ideas, + 5 a competitor would dare” (imposed angles).
→ “Rank in an impact/effort matrix, argued top 3” → you decide.Reading: a full cycle in fifteen minutes.
Lesson: final human decision, with full knowledge.
Section vocabulary
- Divergence
- Produce many ideas without judging: “20 ideas, from classic to wild” — sort afterwards.
Simple example: Amine requests 20 ideas to improve client welcome, from the obvious to the unexpected.
What it is not: asking for a single “good” idea and sticking to it.
Reflex: impose the number and forbid sorting during divergence.
- Imposed angles
- Force variety: “5 low-cost ideas, 5 digital ideas, 5 ideas a competitor would try”.
Simple example: Amine adds 5 zero-budget ideas and 5 ideas his competitor would dare.
What it is not: following up with a plain “give me more” without constraints.
Reflex: impose personas, budgets and constraints at each follow-up.
- Convergence
- Sort with criteria: “Rank these ideas by impact and ease, in a 2×2 matrix”.
Simple example: Amine has his ideas ranked in an impact/effort matrix with an argued top 3.
What it is not: choosing by gut feeling from a scattered list of 30 ideas.
Reflex: require explicit criteria and an argued recommendation.
- Devil's advocate
- “Attack this plan like a competitor / an unhappy client / a CFO”: the ideas stress-test.
Simple example: Amine has his No. 1 idea demolished like a skeptical CFO.
What it is not: asking “it's a good idea, right?” and taking yes for granted.
Reflex: order the critique before any decision.
- Structuring
- From scattered notes to a plan: “Organize these items into a logical 3-part plan with headings”.
Simple example: Amine turns his workshop notes into a 3-part action plan.
What it is not: keeping ideas scattered with no action plan.
Reflex: end each ideation with a titled, actionable plan.
In practice — In practice — A full solo ideation workshop
- Pick a real topic (improve client welcome, lead the team meeting…): “Give me 20 ideas, from the obvious to the unexpected”.
- Force variety: “Add 5 zero-budget ideas and 5 ideas our main competitor would dare”.
- Converge: “Rank everything in an impact/effort matrix and recommend an argued top 3”.
- Stress-test No. 1: “Demolish this idea like a skeptical CFO” — then decide YOURSELF, with full knowledge.
Ready-to-paste script — solo ideation workshop in 4 moves
“Topic: [your topic]. 1) Give me 20 ideas, from the obvious to the unexpected. 2) Add 5 zero-budget ideas and 5 ideas my main competitor would dare. 3) Rank everything in an impact/effort matrix and recommend an argued top 3. 4) Attack your No. 1 idea like a skeptical CFO: 3 major weaknesses. I will decide myself.”
Run on a real topic; keep the final decision — AI proposes, you decide.
Copy-paste template — divergence-convergence instruction
SUBJECT: [___].
DIVERGENCE: 20 ideas (obvious → unexpected) + 5 zero-budget + 5 competitor.
CONVERGENCE: impact/effort matrix + argued top 3.
STRESS-TEST: devil's advocate (skeptic): 3 weaknesses.
DECISION: I decide myself, with full knowledge.Total blank page
Start broad (20 ideas) before narrowing: volume contains the gem.
SME Algiers
Amine prepares 3 tested ideas for his team meeting instead of arriving empty-handed.
Team brainstorming
Human ideas first, then AI to broaden: first, it anchors and constrains the group.
Plan to structure
Scattered notes → logical 3-part plan with headings: the actionable workshop output.
✓ Best practices
- Ask for quantity in divergence (good ideas hide in volume).
- Impose angles and personas to break monotony.
- Have ideas sorted by explicit criteria, then decide yourself.
✗ Common pitfalls
- Letting AI think in your place.
- Stopping at the first ideas (the most banal).
- Forgetting the devil's advocate (the model is complacent).
Key takeaways
- Ask for quantity in divergence: good ideas hide in volume.
- Imposed angles and personas break the monotony of first answers.
- Have ideas sorted by explicit criteria (impact/effort) — then decide yourself.
- The on-demand devil's advocate offsets the model's natural complacency.
48h challenge — apply tomorrow
Run a 15-minute solo workshop on a real topic: 20 ideas, imposed angles, impact/effort matrix, then devil's advocate on No. 1. Note the decision that you make.
- I can run a divergence → convergence → stress-test cycle.
- I can force variety (imposed angles, personas).
- I have decided myself after the devil's advocate.
Frequently asked questions
Are AI ideas banal by design?
The first ones, often yes (it is the statistical average). Value comes from follow-up techniques: imposed angles, constraints, forced combinations — and from YOUR eye spotting the gem. AI is a raw-material generator; sorting and spark remain human.
Can I use AI in a team brainstorming meeting?
Yes, as a second step: human ideas first (so as not to anchor the group), then AI to broaden and challenge. First, it formats the discussion and constrains the group's creativity. Order matters: the human brain diverges better before seeing an “official” list.
What is the right use of AI in ideation?