Learning objectives
By the end, participants can:
- Distinguish task-level AI use from process-level AI workflows
- Evaluate where human oversight should sit within an AI workflow
- Design a multi-step AI workflow for a real piece of work
Workplace outcomes
- A workflow redesigned in your group, with human and AI roles, checkpoints and a success measure
- A scan of your week showing where AI could help, and how much
- First actions committed: what you will start, stop and continue
Learning points
- Three productivity traps, and the slope against the hockey stick
- Eight AI jobs and the human-in-the-loop continuum
- A case study: four fragilities that appear when a pilot meets real conditions
- Three design rules, and GATES for briefing an AI tool
The experience
A design workshop in five steps: recognise, find, test, design and commit. Groups stress-test a case study, then redesign a workflow together.
Every participant keeps an interactive toolkit to use back at work.
Models and frameworks
- Human-AI combinations meta-analysis Vaccaro, Almaatouq and Malone (2024)
- Firm data on AI Yotzov et al. (2026)
This course is part of The Human Advantage Programme. For the full library record, including the course code, use the course library.