TECH 86 — Agentic AI Workflows: Designing Multi-Agent Systems
Quarter: Winter
Instructor(s): Margaret Laffan
Date(s): Feb 6—Feb 7
Class Recording Available: No
Class Meeting Day: Saturday and Sunday
Class Meeting Time: 10:00 am—4:00 pm (PT)
Tuition: $530
Refund Deadline: Jan 30
Unit(s): 1
Status: Registration opens Nov 30, 8:30 am (PT)
What happens when AI agents stop being tools and start becoming co-workers? As agentic AI moves from task automation to coordinated systems of execution, organizations face a challenge beyond which tools to adopt: how to design and manage workflows in which agents work together. A sales team using one agent to draft outreach and another to qualify leads must determine how those agents hand off work and who is accountable when they do not.
Designed for professionals in operational, go-to-market, sales, and business development roles, this course explores what it takes to integrate agents into the way organizations work. Through examples, discussion, and applied exercises, students examine which work belongs with agents, how humans should oversee their decisions, and where judgment remains essential. The course also considers the implications of coordinating multiple agents across a workflow. By the end, students will have a framework for designing agent-enabled workflows and leading effectively as AI becomes part of how work gets done.
Prior exposure to LLMs, prompt design, and agentic AI use cases is required. Basic scripting or low-code experience is recommended. This course will demonstrate third-party AI tools that are not managed or supported by Stanford. Students can explore these tools on their own if they like, but use is optional. Many tools offer free versions or trials; paid subscriptions, if chosen, typically range from $25–$100 per month. See the syllabus for more details.