RESEARCH WORKSPACE · AI SIDE PROJECT

Ear to Mind: From what people say to how they think.

A research workspace for product designers to practice deeper interviews, write cognitive statements, and build mental model maps. I designed the experience around one principle: the researcher keeps the pen. Coaching helps them look again, while the interpretation stays their own.

My role: Product design
Development
Focus: Research workflow
Guided learning
Format: Web app
Interactive demo
Tools: Codex, ChatGPT
React, TypeScript

Challenge: Moving from what someone said to what they are trying to do, without losing the evidence along the way.
Approach: Keep quotes, interpretations, review, and mapping connected in one workspace.
Experience: Six short tutorials introduce the workflow through one fictional research example.

Open the product Try the demo Watch the film

The problem

Designers learning mental models rarely struggle with the diagram. They struggle before it: interviews stop at what people did, and quotes turn into features instead of thinking.

Most AI research tools summarize transcripts for you. Ear to Mind coaches you to do it yourself, using Indi Young’s mental model method.

What they did

“I made the same pasta I always make.”

How they think

“I just didn’t want to make another decision.”

Illustrative quotes from the fictional dinner-planning example.

How it works

01 Interview.

Save the quotes that matter and get neutral follow-up prompts that go one layer deeper.

Interview notes beside neutral follow-up prompts in the fictional dinner-planning demo

02 Statements.

Write the cognitive statement yourself. Get feedback on six checks, never a finished answer.

A researcher drafts a cognitive statement beside the original quote and receives feedback that asks them to reconsider their own wording

03 Map.

Group statements into towers and mental spaces, then find where your product can help.

Cognitive statements grouped in three intention towers across two mental spaces, with support opportunities below

The design

Tutorial 01 · Interviews

Start with what was actually said.

Paste a transcript or import interview notes, keep the speaker labels, and save a selected passage as a quote. The first tutorial introduces the evidence that everything else will return to, using a fictional conversation about deciding what to eat after a long day.

The Interviews tutorial beside the transcript and Save selected quote control — Keep the participant’s words intact before interpreting them.

Tutorial 02 · Saved quote

Keep the source close.

The quote library holds the original passage, the participant, and a link back to the interview. This tutorial follows a saved quote into “Write a statement,” making the transition from evidence to interpretation visible while preserving the words that prompted it.

The Saved quote tutorial showing the original passage and Write a statement action — A statement begins with a traceable piece of evidence.

Tutorial 03 · Statement editor

Write one intention, then look again.

Choose Action, Thought, Feeling, or Philosophy and draft a cognitive statement. The editor keeps the quote beside the researcher’s wording. Coaching offers a strength, a point to reconsider, and a guiding question; the researcher revises the statement and completes six checks before placing it on the map.

The Statement editor tutorial with the source quote, four lenses, review checks and scripted practice feedback — Review the interpretation yourself before it becomes part of the model.

The six checks

Starts with a verb
Describes a process, not a solution
Holds one idea
Fits the chosen lens
Stays faithful to the quote
Uses only the words it needs

Tutorial 04 · Intention tower

Find the intention shared across statements.

Reviewed statements can be grouped into towers using drag and drop or an accessible placement menu. Related towers sit within mental spaces. This tutorial moves from an individual statement to a pattern, with the researcher deciding what belongs together and how to name the shared intention.

The Intention tower tutorial highlighting reviewed statements and their shared intention — Build groups around human intentions, with the source evidence still available.

Tutorial 05 · Support opportunities

Connect possibilities to a human intention.

Support opportunities sit below the line, beneath each intention tower. This tutorial introduces a place to explore how a service or product could help, while keeping those ideas separate from the research above. In the practice example, “Reduce decision effort” leads to making familiar options easier to compare.

The Support opportunities tutorial highlighting the input below an intention tower — Keep what people are trying to do visible while exploring how to help.

Tutorial 06 · Practice coach

Use feedback to deepen your own interpretation.

Try a statement review and return to Coaching history as your thinking develops. Practice feedback is clearly labeled as a scripted example. In a personal project, the researcher prepares a saved review request and uses the connected Ear to Mind plugin in ChatGPT to receive coaching, then brings that perspective back to the workspace.

The Practice coach tutorial with coaching history and clearly labeled scripted feedback — Use questions as a prompt to think again, while keeping authorship of the research.

Reflection

Learning by teaching the machine.

Writing rules for the coach forced me to define what a good cognitive statement actually is. I understood the method better by teaching it to a machine.

Teach at the moment of use.

A small explanation beside the relevant control gives people a next step while leaving room to explore at their own pace.

An independent learning tool informed by Indi Young’s mental model method. Tutorial and demo examples use fictional research data.

Open the product Explore the demo Visit live website