AI-generated Decision Loops for Agentic AI

Identifying Decisions-to-be-Made with Product Context Analyzer

Free trial • No credit card required

What is a Decision Loop?

A Decision Loop connects a decision to the information required to make it and the actions that may follow. Decision Loops are identified for each core task and subtask, making explicit the general knowledge and situational awareness required for decision-making. By making these decision structures visible, they help teams identify opportunities to support users with AI or system intelligence, streamline decision-making, and design interfaces that guide effective choices.

Basic steps to create a List of Decision Loops with Product Context Analyzer

1

Enter a research prompt

Specify the activity you want to analyze. The activity can be a single task, a responsibility, or a multi-phase journey

Example: Order Pizza from a delivery service for a business meeting

Tips

Describe the activity in a single sentence without prescribing a particular product or solution.

Optionally include context such as who is performing the activity and under what circumstances.

Adding context helps make the analysis more domain-specific.

Activities involving judgment, planning, or coordination provide particularly useful input for decision analysis

2

Click 'Start Analysis'

The analysis takes several minutes depending on the scope and number of tasks identified

3

Select a Task

Select a task from the task hierarchy to explore its detailed context-of-use information

Tips

Decision Loops are identified for the selected task and its subtasks

4

Select the 'Decision Loops' view

Select the Decision Loops view to see the decisions identified across the selected task and its subtasks

Tips

Each Decision Loop shows the required information, the decision to be made, and the actions to be taken

Required information includes both general knowledge and situational awareness

Use Decision Loops to identify where AI can provide information, support decisions, or take action

Use Decision Loops together with Jobs-to-be-Done as inputs for defining AI agent responsibilities

What can you do with Decision Loops?

Run an AI opportunity workshop

Walk through the decisions in a task flow to systematically identify where AI could provide information, support a decision, recommend an action, or act on the user's behalf.

Decide what the AI should know

Use the required information for each decision to identify the general knowledge and situational awareness an AI capability needs to provide meaningful support.

Decide what the AI should do

Use the actions associated with each decision to explore where AI should advise the user, prepare an action, or execute it autonomously.

Define an AI agent's responsibilities

Combine recurring decisions, required information, and possible actions across tasks to define a coherent responsibility for an AI agent.

Combine it with Jobs-to-be-Done

Use desired outcomes to define what an agent should help accomplish and Decision Loops to determine where and how intelligence can help accomplish it.

External Resources

Origins

Joerg Beringer, Alexander-John Karran, Constantinos K. Coursaris & Pierre-Majorique Léger, HCII 2023

Established Practice

Joerg Beringer, Alexander J. Karran, Constantinos K. Coursaris & Pierre-Majorique Léger, 2025

Contemporary Thinking

PCA Perspectives

Product Context Analyzer
Product Context Analyzer
ProContext Inc.
377 Roble Ave
Redwood City
CA 94061, USA