Jeff Camm spent 30 years at the University of Cincinnati before moving to Wake Forest to build its master's in business analytics from scratch. He argues that many analytics projects begin in the wrong place. Organizations collect large amounts of data and then look for something useful to do with it. Camm starts by asking which decision needs improvement, what constraints shape it, and what evidence would help someone make that choice.
Camm and his colleagues studied how employers and universities define the skills needed for data science and decision science. Across roughly 1.4 million data science job ads and 50,000 decision science ads over ten years, they found that data science roles emphasized coding, machine learning, and big data, while decision science roles more often mentioned business domains and problem solving. In the master’s programs they studied, required data science coursework devoted about 28% of credit hours to prediction and 2% to decision modeling. Camm sees value in both skill sets, but argues that a prediction creates business value only when someone can use it to choose an action.
That view shaped the business analytics program he built at Wake Forest. Students learned technical methods, but the curriculum also focused on two areas where projects often fail: framing the business problem and ensuring a recommendation is used. A practicum asked students to work from an unclear business situation toward a model, while a course combining visualization and influencing taught them how to explain findings to decision-makers. Camm says employers noticed that graduates could both communicate recommendations and conduct analysis.
He now proposes a similar sequence for undergraduate business education. An AI course would teach students to ask better questions, supply context, verify outputs, and use the tools responsibly. Business intelligence would help them understand operations through descriptive and predictive analysis. Decision intelligence would bring in optimization, simulation, uncertainty, behavioral factors, and change management. His point is that students should learn each method in the context of a business choice, including the costs of being wrong.
Camm also changed how he presents the optimization results. Earlier in his consulting career, he gave clients a model’s best answer. He later began showing them several feasible solutions so they could weigh risks and considerations the model had not captured. He applies the same idea to AI. Ask for alternatives, examine their costs and risks, and leave the final trade-off with the person responsible for the decision. For analytics leaders, the test of a project is whether it changes a decision and whether the organization can explain why.
The DecideWise Insider
Benjamin Baer, Founding Member of the DecideWise community, argues that decision velocity depends on measuring the quality and impact of decisions alongside their speed. Processing volume and system uptime indicate whether an automated process is running, but they do not indicate whether its decisions improve business outcomes or remain within risk limits. Baer proposes tracking decision cycle time, the share of decisions completed without manual intervention, errors and overrides, and the value gained or lost against alternative choices. He argues that teams should define these measures before deployment and build outcome tracking into the decision process. That gives domain experts evidence to spot drift, review exceptions, and adjust rules when conditions change.
Weigh in, your feedback starts with one quick vote.
What do you think is the future of AI evolution?
Market Pulse
Sapiens launched its Autonomous Insurance Platform, adding AI workflows for underwriting, policies, billing, and claims. Its Migration Hub proposes how legacy data should map into a new system and gives each suggestion a confidence score for insurance staff to review. Its Configuration Hub maps requirements documents to system fields, lets staff approve the proposed changes, and keeps an audit trail of each decision. Continental General is evaluating both tools to help bring new books of business onto its systems.
Experian launched Activate to match consumers with credit offers based on lenders' underwriting criteria and information available during the shopping session. Participating lenders can combine credit data with cash flow insights from bank accounts consumers choose to connect, allowing them to assess recent income and banking activity alongside credit history. The decision engine serves Experian Marketplace, which has more than 90 million members.
IBM announced a private preview of Agent Identity in watsonx Orchestrate. It assigns each agent a distinct identity and preserves both the agent and the requesting user in the execution record. When an agent calls a protected tool, the authorization system can evaluate that relationship and grant access for the specific task. The resulting audit record links the user, agent, authorization decision, credential, and tool action. The preview supports integrations with IBM Verify and Microsoft Entra.
Alation announced six additions to its AIOS platform, including tools to track AI agents and define the business context they use. AI Governance links an agent’s lineage and regulatory risk to the data it consumes, with connectors for six model and agent platforms. Governed Collections gives agents access to current policies and operating procedures, while Ontologies makes business rules and constraints explicit for agents to use. The governance and semantic model updates are available now, and the other four products are in early access.
Resources and Events
📅 CDAO Dallas 2026 (Dallas, TX - December 2, 2026)
CDAO Dallas returns on December 2 at The Highland Dallas for a one-day program aimed at chief data and analytics officers, CIOs, and other senior data leaders. The 2026 themes cover data governance and compliance, business applications of AI, analytics for forecasting and operational decisions, and the ethics of AI use. The event provides leaders from different industries with a forum to compare how they set priorities for data initiatives, measure their results, and manage the risks that come with AI use.
📅 Big Data Conference Europe 2026 (Vilnius, Lithuania - November 24-27, 2026)
Big Data Conference Europe takes place November 24-27 in Vilnius, Lithuania, with online access to the three main conference days. The first day is devoted to workshops, followed by a program of more than 90 planned talks across six tracks. Topics include data engineering, streaming architectures, cloud platforms, analytics, machine learning, data security, and AI agents. Speakers include leaders from AWS, Databricks, Exacaster, and Perplexity.
📊 Report Spotlight: The AI Agent Management Gap (Guild.ai)
This report, based on a Morning Consult survey of 362 U.S. IT decision-makers, finds that 96.4% are confident their AI agent inventory is complete, but only 42.7% have a centralized monitoring tool and 39.8% have audit logs. Among organizations with agents, 66.7% reported an agent-related operational consequence in the past year. While 60.6% think employees may deploy agents without approval, only 35.9% require registration before deployment.
Join the conversation: When does optimizing one metric distort the decision?
Join the DecideWise discussion on how narrow targets can hide effects elsewhere in a system, why teams keep relying on them, and what practical steps could help decision-makers examine the wider consequences.
For the Commute
Business Lessons from Sports Analytics (Decision Intelligence Lab)
Vijay Mehrotra and Michael Watson speak with Ben Alamar about the evolution of sports analytics and the difficulty of getting evidence used in consequential decisions. Alamar has worked in basketball operations with the Oklahoma City Thunder and Cleveland Cavaliers, led sports analytics at ESPN, and later held analytics roles at StubHub and Apple Ads. He discusses how motion tracking expanded what teams can measure, and how coaches and executives interpret that data when evaluating players, making draft picks, and weighing risk. The conversation also covers leadership’s role in building an evidence-based culture, how lessons from sports analytics carry over into business decisions, and advice for people entering the field.
