For decades, supply chain organizations adapted their operations to the limitations of their technology. Employees filled gaps between systems, while organizational structures were designed around what existing software could support. Physical supply chains often moved faster than companies could collect and interpret information about them. Sasha Koff, a supply chain executive with more than two decades of experience at Dell and Johnson & Johnson, believes this relationship has reversed. Connected technologies can now anticipate events faster than physical supply chains can respond. Consequently, expanding AI and decision intelligence depends on whether companies can redesign processes, develop employees, and coordinate decisions across functions.
Koff identifies executive sponsorship as a prerequisite for transformation. Employees must recognize their authority, and the sponsor must control sufficient funding to establish teams, acquire technology, and redistribute resources. They also need credibility among other executives because supply chain decisions affect procurement, manufacturing, finance, and other departments. A transformation program may struggle even with strong supply chain leadership if its sponsor cannot secure cooperation from those functions.
Leaders must be willing to experiment and must understand the outcome they want and the processes required to achieve it. Expecting technology to determine those objectives is a recurring cause of failure. Organizations also need employees who understand their data, existing architecture, and the connections between systems, including technologies supplied by external partners.
The next challenge is developing leaders who understand both supply chain operations and the technical systems supporting them. Operational knowledge allows leaders to understand business constraints and supply chain decisions. Technical knowledge enables them to examine analytical models, question assumptions, understand algorithms, and assess whether recommendations are feasible. Leaders do not need to build every model themselves, but they must understand enough to evaluate their teams and set appropriate boundaries, governance, and organizational structures as software takes on a greater role in decision-making. Finding people with both capabilities is difficult because supply chain and technology careers have traditionally developed along separate paths.
External partnerships require a clear understanding of what a company can do internally and where it lacks expertise. Vendors typically provide additional capacity for activities the company already performs well, while partners bring capabilities it does not possess. For example, a supply chain team may have the necessary data but lack the analytical models to use it effectively. A technology partner can fill that gap, provided both sides understand their responsibilities and can rely on each other's expertise. However, external support does not transfer accountability. Executives remain responsible for the program's decisions, technology, data, and results, even when specialists handle implementation. They must understand the systems being developed, question the underlying assumptions, and ensure the technology delivers the intended outcomes.
The DecideWise Insider
Benjamin Baer, Founding Member of the DecideWise community, argues that decision systems need continuous improvement after deployment, since any decision model, rule set, or machine learning framework can be refined over time. He asks teams to record the rationale and expected metrics behind every change, then return months later to check the results. Decision logic drifts as much as machine learning models do, and he argues leaders should realign it with business goals on a regular cycle. Simulation on historical data, synthetic datasets, or stress scenarios lets teams test new models, failure modes, edge cases, and guardrails before production. He sees the larger gains in simulating connected decisions, since a pricing change moves through inventory allocation, finance risk limits, and support capacity.
Market Pulse
Workday has expanded its Decision Intelligence platform to HR and finance, allowing teams to analyze workforce and financial data, compare scenarios, and make decisions within existing approval processes. HR teams can evaluate hiring plans, skills gaps, employee retention, and staffing costs, while finance teams can investigate revenue changes, margin declines, and budget variances. The system uses live Workday data and maintains records of assumptions, approvals, and decisions. The capabilities are now available to Workday HCM and Financial Management customers.
Predictiv AI has launched Flockentis, an AI-powered planning system that evaluates fleet costs, delivery schedules, driver availability, regulatory requirements, and carbon emissions before dispatch. The platform can process up to 5,000 jobs per planning cycle and generate three alternative operating plans based on cost, service performance, or environmental impact. It also identifies deliveries that cannot be completed within available resources or operating constraints.
Microsoft has launched Microsoft-Decision-1, a model designed for classification, prioritization, routing, verification, and controlling AI agent actions. Built on Qwen3.5-9B, it assigns probability scores to predefined options. In Microsoft's evaluation across 36 benchmarks and nearly 150,000 questions, the model achieved the highest accuracy and was 35 times faster than GPT-6 Sol. Internal testing also found it could classify more than 10,000 customer feedback entries at 200 times lower cost. The model is available through Microsoft Foundry and OpenRouter, with input pricing of $0.042 per million tokens.
H2O.ai's Lightning-4B decision model ranked first among 104 open-weight systems on the JevBench leaderboard, scoring 72.5 out of 100. Built on Qwen3.5-4B, the model handles classification, yes-or-no questions, and scoring tasks and returns probabilities. It recorded a median latency of 29 milliseconds per decision and an estimated cost of $0.021 per 1,000 decisions. JevBench evaluates accuracy, probability calibration, speed, and cost equally. Released under the Apache 2.0 license, the model can run on-premises, in private clouds, or in isolated environments.
Resources and Events
📅 INFORMS Annual Meeting 2026 (San Francisco, CA - November 1-4, 2026)
The meeting takes place at Moscone Center South and the San Francisco Marriott Marquis, bringing together more than 6,000 researchers, industry experts, students, and employers to explore operations research and analytics. The in-person program includes technical presentations, poster sessions, technology showcases, and exhibitor workshops, with a dedicated AI poster session on November 2. A career fair and community events offer opportunities to recruit specialized talent and connect with researchers and practitioners.
📅 The Agentic Transformation Summit (Amsterdam, Netherlands - November 4-5, 2026)
The Agentic Transformation Summit brings together more than 200 senior enterprise leaders to examine how companies can deploy autonomous AI agents at scale. Held alongside the Business Transformation Europe Summit, the event focuses on integrating AI agents into back-office operations, modernizing enterprise architecture, measuring financial returns, and establishing governance practices aligned with the EU AI Act. More than 40 speakers from organizations including Mastercard, Santander, IKEA, and Philips will discuss practical challenges in deploying AI agents, managing organizational change, and demonstrating measurable business results.
📅 Data & Analytics Summit (Chicago, IL - October 29, 2026)
This one-day summit at The Ritz-Carlton Chicago focuses on how consumer goods companies apply data, analytics, and AI to commercial and operational decision-making in the US market. Themes include turning data into better business decisions, analytics that drive commercial growth, scaling AI across the enterprise, and building high-performing analytics organizations. Speakers include analytics, insights, and revenue growth leaders from Hershey, Kraft Heinz, Kenvue, Ahold Delhaize, Conagra, General Mills, and Costco.
📊 Report Spotlight: State of Responsible AI 2026 (FICO)
FICO's 2026 global survey of AI and technology leaders finds that only 5.2% of respondents are very confident in their ability to explain AI decisions to regulators and customers. Just 8.2% of organizations have fully implemented a shared AI deployment platform, while agentic AI remains largely confined to pilot projects amid security and data concerns. AI currently handles only 28.5% of customer-impacting decisions, with the remainder relying on employees, rules-based systems, and traditional analytics.
Join the conversation: What is the future of AI evolution?
Join the DecideWise discussion on the future of AI, from seamless integration into everyday life to job losses, loss of control, and the possibility of widespread societal disruption. Share your perspective on where AI development is heading.
For the Commute
From Dynamic Forecasting to Perfect Order (The Decision Intelligence Podcast)
Luis Gonzalez, Global Supply Chain Director at Becle, the company behind Jose Cuervo, discusses how decision intelligence is changing forecasting, inventory management, and order execution across its global operations. Becle expanded dynamic forecasting from its largest market to 92 countries, reducing forecast error by 50% in the first year. The company is now integrating Aera with SAP through its Perfect Order initiative to automate routine supply chain decisions. Approximately 60% of these decisions can already be made autonomously, with human oversight maintained as automation expands. Gonzalez also examines how companies must adapt their organizational structures to make decisions at the speed newer technologies allow.
