Universities are moving from delivering content to cultivating thinking—where students, instructors, and personal AI agents (PAIAs) co-construct understanding in real time. Building on two decades of blended learning, the instructor shifts from “source of answers” to “designer of inquiry,” orchestrating interactions among learners, scholarly resources, and agentic tools. We propose a digital intelligence framework that operationalizes this shift through three components: (1) a Reading-to-Questions (R2Q) pipeline that ingests texts, distills claims, cross-checks sources, and generates tiered Socratic prompts; (2) agent-to-agent seminar orchestration in which student and instructor PAIAs negotiate agendas and produce “conflict maps” to surface productive disagreements; and (3) an AI-augmented blended layer aligning online and in-person activities with human–agent partnerships for preparation, discussion, and transfer tasks. Our primary learning KPI is the quality of thinking—captured via reflective depth, reasoning robustness, question quality, cross-domain transfer, and synthesis velocity—tracked longitudinally at individual and cohort levels. The approach turns reading into collaborative inquiry, scales rigorous dialogue, and prepares graduates for hybrid human–AI work while honoring the classic insight that learning and thinking must reinforce one another.