Decision Intelligence Platform
Gartner designated Decision Intelligence "transformational" in the 2025 AI Hype Cycle and published the inaugural Magic Quadrant for Decision Intelligence Platforms in January 2026. Autonomy is the first purpose-built DI platform for supply chain.
What Is Decision Intelligence?
Gartner defines Decision Intelligence as "a practical discipline that advances decision making by explicitly understanding and engineering how decisions are made, and how outcomes are evaluated, managed and improved via feedback." By digitizing and modeling decisions as assets, DI bridges the insight-to-action gap.
Google's first Chief Decision Scientist frames it as "the discipline of turning information into better actions at any setting, at any scale" — unifying three sub-disciplines that traditionally operate in silos:
- Applied Data Science — Analytics, statistics, ML for extracting insights
- Social Science — Group dynamics, stakeholder perspectives, cognitive bias
- Managerial Science — Goal alignment, process management, organizational structure
Gartner defines Decision Intelligence Platforms as software that supports, automates, and augments decision-making by bringing together data, analytics, knowledge, and AI — while enabling collaboration across decision modeling, execution, monitoring, and governance.
The Decision Lifecycle
Gartner's four critical Decision Intelligence capabilities, implemented natively for supply chain in Autonomy.
Decision Modeling
Define what decisions exist and how they work
A sequential decision framework models every decision with five elements: State (inventory, backlog, pipeline), Decision (order quantity, allocation), Exogenous Information (demand, lead times), Transition Function, and Objective (dollar-denominated outcomes).
Implemented: 11 agent definitions, state decomposition across physical, information, and belief dimensions
Decision Orchestration
Coordinate execution flows across agents and systems
A 6-phase decision cycle coordinates 11 agents per site. The Agentic Authorization Protocol (AAP) handles cross-functional trade-offs at machine speed — 25+ negotiation scenarios across manufacturing, distribution, procurement, and finance.
Implemented: Signal propagation, urgency coordination, authorization surfaces
Decision Monitoring
Track outcomes, detect drift, measure quality
Hourly outcome collection scores every decision against actual results. Calibrated confidence intervals track certainty. Forecast quality scoring measures probabilistic accuracy. Drift triggers detect when agents need retraining.
Implemented: Outcome collection, confidence calibration, quality scoring, drift detection
Decision Governance
Ensure compliance, auditability, and trustworthiness
Override effectiveness is tracked statistically per user and decision type. Confidence bounds on every decision provide mathematical guarantees. The Escalation Arbiter routes persistent drift to higher reasoning tiers. Full audit trail from decision to outcome.
Implemented: Override tracking, authority boundaries, escalation log
Three-Level Maturity Progression
Autonomy's progression from Support to Automation is governed by measured decision quality — not arbitrary trust thresholds or time-based milestones.
Decision Support
Human in the loop. The system provides data, insights, scenarios, and reports. All decisions require human input. This is traditional BI and planning software.
Decision Augmentation (Copilot)
Human on the loop. AI agents generate recommendations with impact analysis. Humans approve, modify, or reject. Every override is captured with reasoning and scored against outcomes. This is where most customers start.
Decision Automation (Autonomous)
Human out of the loop. AI agents execute autonomously within pre-approved guardrails. Full auditability. Humans focus on governance, exception review, and strategic decisions. Progression governed by calibrated confidence, override effectiveness scores, and decision quality metrics.
Decisions as Digital Assets
In traditional planning software, decisions are implicit — they're the output of a planning run, buried in a supply plan or MPS schedule. You can see the result, but not the decision that produced it, why it was made, what alternatives existed, or what happened afterward.
Autonomy treats every recurring supply chain decision as a first-class digital asset with:
- Defined inputs and triggers — What state caused this decision?
- Explicit logic and constraints — What model produced it? What guardrails apply?
- Clear ownership and authority — Which agent type? What authority boundary?
- Measurable outcomes linked to actions — What actually happened?
- Feedback loops for continuous improvement — How does this outcome improve future decisions?
This transforms planning from "generate a plan and hope it works" to "make thousands of tracked, governed, improving decisions every day."
Causal Decision Diagrams
Every agent maps to a formal Causal Decision Diagram (CDD):
- Decision Levers — Actions the agent can take (order quantity, transfer amount, release/defer)
- Externals — Factors outside control (demand variability, lead time uncertainty, supplier reliability)
- Intermediaries — Leading indicators along the causal chain (fill rate, days of supply, pipeline position)
- Outcomes — Ultimate measurable goals (total cost, OTIF, inventory turns)
The CDD makes the causal chain transparent: how does this specific order quantity decision flow through inventory levels, service rates, and ultimately to dollars saved or lost?
Market Context
of SCM solutions will use intelligent agents by 2030 (Gartner)
of Global 500 will apply DI practices including decision logging by 2026 (Gartner CDAO Survey)
Inaugural Gartner Magic Quadrant for Decision Intelligence Platforms published
25% of CDAO vision statements will become "decision-centric" surpassing "data-driven" (Gartner)
Four Interdependent Technologies
Gartner's 2025 Hype Cycle for Supply Chain Planning identifies four technologies as "interdependent levers of change, not individual trends." Autonomy implements all four natively.
Decision-Centric Planning
Innovation Trigger
The organizing principle. Shift from periodic batch planning to continuous decision execution. Every recurring choice modeled as a repeatable decision asset.
Agentic AI
Innovation Trigger
Autonomous agents specialized in different areas interact seamlessly, creating integrated supply chain views. Autonomy deploys 11 per site today.
Autonomous Planning
Trough → Slope of Enlightenment
Demands a cultural shift from people-centric to decision-centric. Measured maturity progression governs the transition.
Explainable AI
Slope of Enlightenment
Proven but underused. Every Autonomy agent decision includes reasoning with evidence citations and confidence scores. Ask Why at any point.
Autonomy vs. Horizontal DI Platforms
Gartner's MQ Leaders (SAS, FICO, Aera Technology) are horizontal platforms that require extensive customization for supply chain.
| Capability | Horizontal DIPs | Autonomy |
|---|---|---|
| Decision Modeling | Generic business rules | Domain-specific sequential decision framework for supply chain |
| Decision Execution | Rules engines, workflow | Real-time specialized agents (<10ms) |
| Decision Monitoring | BI dashboards | Calibrated confidence + quality scoring + drift triggers |
| Decision Governance | Audit logs | Statistical override tracking + causal inference |
| Supply Chain Domain | Bolt-on or absent | Native (35 AWS SC entities, 8 policy types) |
| Agentic AI | Early / experimental | 11 production agents per site, multi-site coordination |
| Probabilistic Planning | Limited | 21 distributions, Monte Carlo, forecast quality scoring |
| Learning from Overrides | Basic | Statistical tracking + causal forests |
See Decision Intelligence in action
Watch how decisions flow from modeling through execution to measured outcomes.