Un Prompt de matrice de décision pondérée pour trancher plus vite en équipe
Why this prompt matters
Teams waste weeks revisiting the same decision when criteria are implicit, stakeholders optimize for different goals, and nobody can explain why an option won. A weighted matrix turns vague debate into an auditable decision record and exposes where disagreement actually lives.
What we use it for
Use this when you need to choose between several real options under time pressure, such as selecting a project management tool, vendor, framework, or job offer, and you want a defensible recommendation instead of gut feel.
Prompt
Role: Act as a senior strategy analyst helping me make a high-stakes decision with a transparent weighted scoring model. Context: I am deciding between [OPTION 1], [OPTION 2], [OPTION 3], and [OPTION 4] for [DECISION CONTEXT]. The decision matters because [WHY THIS DECISION MATTERS]. My constraints include [BUDGET], [TIMELINE], [TEAM SIZE], and [NON-NEGOTIABLE REQUIREMENTS]. Task: Build a weighted decision matrix that compares the options across the criteria I provide. Use these criteria and weights, adding up to 100: [CRITERION 1 + WEIGHT], [CRITERION 2 + WEIGHT], [CRITERION 3 + WEIGHT], [CRITERION 4 + WEIGHT], [CRITERION 5 + WEIGHT]. Score each option from 1 to 10 on every criterion, explain each score in one or two sentences, calculate weighted totals, rank the options, and recommend the best choice. Then stress-test the recommendation by showing which option would win if the top-weighted criterion changed by plus or minus [SENSITIVITY %]. Constraints: Do not use vague language. If information is missing, list the missing inputs explicitly and make the smallest reasonable assumption. Do not force a tie. Call out hidden risks, second-order effects, and cases where the highest-scoring option may still be the wrong practical choice. Output Format: 1) Decision summary, 2) assumptions and missing data, 3) weighted decision matrix in table form, 4) ranking with weighted totals, 5) recommendation with rationale, 6) sensitivity analysis, 7) next steps before committing.
Result
Decision summary: Based on the weighted matrix, Notion is the strongest fit for a 12-person remote product team choosing a project management platform for the next 18 months. It does not win every category, but it delivers the best overall balance of flexibility, cross-functional collaboration, documentation, and cost. Jira remains the better pure engineering tracker, while Asana is the easiest to adopt quickly, but neither matches Notion's combined score across the broader operating model this team needs.
Assumptions and missing data: I assumed the team needs one shared workspace for product, engineering, design, and operations, with a monthly software budget below $500. I also assumed the company values fast onboarding and documentation quality as much as sprint planning. Missing inputs that would affect the outcome include existing tool lock-in, compliance requirements, and whether the engineering team needs advanced issue hierarchy and release reporting.
Weighted decision matrix: Criteria were weighted as follows: collaboration and documentation 30, ease of adoption 20, engineering workflow support 20, automation and integrations 15, total cost 15. Notion scored 9, 8, 7, 8, and 9 respectively, producing a weighted total of 8.30. Jira scored 6, 5, 10, 9, and 7 for a total of 7.20. Asana scored 8, 9, 6, 7, and 6 for a total of 7.35. ClickUp scored 7, 6, 7, 8, and 8 for a total of 7.10.
Recommendation with rationale: Choose Notion if the company wants one system that supports planning, documentation, meeting notes, lightweight roadmapping, and cross-team visibility without adding another knowledge base. Its main weakness is deeper engineering workflow structure, so teams with strict release governance may still prefer Jira despite the lower blended score.
Sensitivity analysis: If engineering workflow support rises from 20 to 35, Jira becomes much more competitive and may overtake Notion, especially if the company already uses GitHub and wants stronger issue discipline. If collaboration and documentation fall below 20, Asana also narrows the gap. That means the decision is robust only if the organization truly values a shared operating system over a specialist tracker.
Next steps before committing: Run a two-week pilot with one live project, measure onboarding time, check migration friction, and ask each function to list one blocker that would make the chosen tool fail in practice.
Choisir entre plusieurs bonnes options est l'un des endroits où les équipes perdent le plus de temps. La discussion semble utile, mais les critères restent implicites, la voix la plus forte pèse sur le résultat, puis la décision est rouverte deux semaines plus tard. Ce Prompt corrige cela en forçant les arbitrages dans une matrice pondérée que toute l'équipe peut examiner.
La structure compte. La section Role place le modèle en mode analyste plutôt qu'en brainstorming générique. La section Context définit les enjeux, les contraintes et les éléments non négociables, afin que la recommandation reste ancrée dans une vraie décision métier au lieu d'une préférence abstraite. La section Task fait le travail principal: noter chaque option, expliquer chaque note, calculer les totaux pondérés, puis tester la robustesse du résultat avec une sensitivity analysis.
C'est cette dernière partie qui rend ce Prompt vraiment utile à conserver. Un Prompt de comparaison classique vous donne une liste classée. Un meilleur Prompt montre si le gagnant reste gagnant quand le critère le plus important varie. Si un petit changement de pondération inverse le résultat, le vrai problème n'est ni l'outil ni le Vendor. C'est l'alignement entre les parties prenantes.
Ce Prompt évite aussi un échec fréquent de l'AI: des réponses très assurées construites sur des informations manquantes. La section Constraints demande au modèle d'indiquer les données absentes, de faire l'hypothèse raisonnable la plus minimale possible et de signaler les cas où l'option la mieux notée n'est pas forcément le meilleur choix pratique. On obtient ainsi quelque chose de plus proche d'un decision memo que d'une supposition élégante.
Utilisez-le pour choisir un logiciel, préparer une décision de recrutement, évaluer une agence, comparer des modèles tarifaires, arbitrer une roadmap ou toute autre décision où plusieurs bonnes options se disputent des critères différents. Le résultat se copie facilement dans un document, se défend en réunion et peut être réévalué plus tard si les hypothèses changent.