Curriculum

Cluster 4 ¡ Lesson 4 1 min read

Where Does AI Fit?

Identifying which steps could benefit from AI assistance and which should not.

Integrating AI into your workflow is not an all-or-nothing proposition. The most effective professionals understand that AI is a tool with specific affordances, not a universal solution. In this lesson, we explore where AI fits best by applying the IDEO affordances framework, which categorizes AI assistance into four distinct modes: Accelerate+Expand, Accelerate+Interrogate, Deliberate+Expand, and Deliberate+Interrogate.

Equally important is recognizing where AI does not belong. Tasks requiring human judgment, emotional intelligence, and ethical decision-making must remain human. Furthermore, because AI capabilities form a "Jagged Frontier"—excelling at some complex tasks while failing at simpler ones—you cannot simply predict its utility. You must test it. This lesson will guide you in critically evaluating your workflow to determine exactly where AI can add value and where human insight is irreplaceable.

Assignment

Review the workflow you mapped in the previous lesson. For each step, apply the IDEO affordances framework. Label steps as Accelerate+Expand, Accelerate+Interrogate, Deliberate+Expand, Deliberate+Interrogate, or Human-Only. Write a brief justification for any step you labeled as Human-Only.

Learning Objectives

  • Understand that not every step in a workflow requires or benefits from AI.
  • Apply the IDEO affordances framework to identify the right mode of AI assistance.
  • Embrace the concept of the Jagged Frontier by testing AI capabilities rather than predicting them.

The IDEO Affordances Framework

A model for identifying how AI can assist in different modes: Accelerate+Expand (generating many options fast), Accelerate+Interrogate (stress-testing ideas), Deliberate+Expand (deep pattern finding), and Deliberate+Interrogate (careful analysis).

The Jagged Frontier

The unpredictable boundary of AI capabilities. Because AI performance varies wildly across seemingly similar tasks, you cannot assume what it can or cannot do—you must test it empirically.

The binding constraint on AI adoption is not technical skill, but the ability to see your own work clearly enough to redesign it.