Decision intelligence is the discipline of applying data science, AI, and behavioral science to improve decision-making at every level of an organization.
It sits above traditional analytics and below pure AI. It's not just about showing data (that's BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition →). It's not just about building models (that's data science). It's about understanding how decisions are actually made, where they fail, and how data and models can improve them.
Organizations invest heavily in analytics but often see limited improvement in decision quality. Why?
The knowing-doing gap: Data is available but not consulted. Managers make intuitive decisions and look for data to confirm them afterward. This is confirmation bias, and it's endemic.
Decision friction: The data that would improve a decision is hard to access or requires interpretation. If accessing the right data takes 30 minutes, people make the decision without it.
Cognitive overload: Too many dashboards, too many metrics. Analysis paralysis is a real cost, decisions get delayed waiting for more data.
Decision intelligence addresses these root causes, not just the data supply.
Knowledge check
1. According to the lesson, how is decision intelligence positioned relative to other disciplines?
2. Which type of decision in the Decision Architecture Framework is characterized as infrequent, high-stakes, and hard to reverse?
3. The lesson references 'confirmation bias' as part of the knowing-doing gap. In a decision intelligence context, what is the most effective countermeasure?
4. Select ALL the root causes that explain why organizational decisions fail according to the lesson.
Sélectionnez toutes les réponses correctes.
5. Select ALL correct statements about operational and automated decisions in the framework.
Sélectionnez toutes les réponses correctes.
Decision intelligence starts with mapping the decision landscape:
Strategic decisions: Infrequent, high-stakes, hard to reverse. Examples: market entry, M&A, major technology investment. Data role: scenario modeling, market analysis, risk quantification. Human judgment dominates.
Tactical decisions: Weekly/monthly, moderate stakes. Examples: budget allocation, product prioritization, pricing. Data role: trend analysis, A/B testA/B testA/B testing is a controlled experiment that compares two versions of something (A and B) by splitting traffic randomly to learn which performs better on a chosen metric.View full definition → results, forecasts. Data-informed human decisions.
Operational decisions: Daily/hourly, lower stakes, high volume. Examples: inventory reorder, ad bid management, fraud flag review. Data role: automated rules and models. Human oversight at scale.
Automated decisions: Millisecond, very high volume. Examples: real-time fraud scoring, content personalization, dynamic pricingdynamic pricingAutomatically adjusting prices in real time based on demand, competition or user behaviour to optimise revenue, margin or conversion.View full definition →. Data role: ML models making autonomous decisions. Human monitors for drift.
Most organizations invest heavily in strategic analytics and neglect the operational and automated tiers, where decision volume is highest and where automation delivers the most ROIROIReturn on Investment: the ratio of net profit to the cost of an investment. A 300% ROI means each dollar invested returns $3.View full definition →.
A/B testingA/B testingA/B testing is a controlled experiment that compares two versions of something (A and B) by splitting traffic randomly to learn which performs better on a chosen metric.View full definition → is the gold standard for causal inference in product and marketing decisions. But most organizations don't have the infrastructure to run rigorous tests.
A mature A/B testingA/B testingA/B testing is a controlled experiment that compares two versions of something (A and B) by splitting traffic randomly to learn which performs better on a chosen metric.View full definition → program requires: a randomization framework (ensure users are randomly assigned, not self-selected), sufficient sample sizes (understanding statistical power before running tests), multiple comparison correction (running 20 tests simultaneously guarantees false positives), and clear decision rules (what statistical confidence level triggers a decision to ship?).
Booking.com runs over 1,000 simultaneous A/B tests. Every product change is tested before full rollout. The culture: intuition forms hypotheses, data decides which hypotheses are correct. This is decision intelligence at scale.
The most impactful analytics is invisible: data embedded in the workflow where the decision is made, not in a separate dashboard someone has to remember to check.
Examples:
Building this requires integration between the data platform and operational tools, technical complexity that many organizations underestimate.
1. Quelle est la principale différence entre la Business IntelligenceBusiness IntelligenceTechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → traditionnelle et la Decision Intelligence ?
A) La BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → utilise des données historiques, la DI uniquement des données temps réel
B) La BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → montre des données, la DI comprend comment les décisions sont prises et intègre données et modèles pour les améliorer
C) La Decision Intelligence remplace les analystes humains
D) La BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → est moins coûteuse que la DI
Réponse: B
2. Pour quel type de décision l'automatisation par ML offre-t-elle le ROIROIReturn on Investment: the ratio of net profit to the cost of an investment. A 300% ROI means each dollar invested returns $3.View full definition → le plus élevé ?
A) Les décisions stratégiques de haut niveau
B) Les décisions tactiques mensuelles
C) Les décisions opérationnelles à très haut volume comme la détection de fraude ou la personnalisation
D) Les décisions de M&A
Réponse: C
3. Pourquoi les décisions échouent-elles souvent malgré la disponibilité des données ?
A) Les données sont de mauvaise qualité
B) Les outils BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → sont trop complexes
C) Le "knowing-doing gap" : confirmation bias, friction d'accès, surcharge cognitive et désalignement des métriques
D) Les équipes n'ont pas accès aux outils analytiques
Réponse: C