Our Seamless Stories workflow in practice

Automating chat interviews with Qualia. Then using Causal Map to make sense of them. A case study from Chile.

Our seamless AI-supported workflow is coming together. We recently helped colleagues at a university in Chile to complete a qualitative, exploratory evaluation, using Qualia to conduct the interviews and Causal Map to analyse them.

This workflow lets you do in-depth research more quickly and cheaply than before while maintaining depth and quality.

Background

DuocUC, a higher education institution in Chile, commissioned QuIP-style interviews with Qualia and analysis in the Causal Map app. The study sat in the quality assurance department, led by Felipe Rivera, Head of Academic Quality Evaluation. It was motivated by concerns about gender gaps for women pursuing STEM careers.

After scoping calls to agree research questions and domains, we drafted Qualia’s instructions for the interview, iterating with the client’s team.

Step 1: Setting up the interview in Qualia

• Instructions for the AI interviewer were similar to those you would give a human.
• Both the instructions and the interviews themselves were in Spanish.
• The AI asked about changes in three domains: educational experiences, professional development, and relationship dynamics.
• Model: GPT-4o.

Step 2: Collecting stories

• We sent the interview link to 50 people and collected 32 interviews.
• Personalised keys (e.g. &key=0003) at the end of each invitation let the researchers track who responded without storing names in Qualia.
• We downloaded the results and uploaded them into Causal Map.

Step 3: Analysing with Causal Map

• AI (GPT-4o) identified each causal link in the interviews, labelling cause and effect.
• We used a “radical zero-shot” approach: no codebook, the AI invented its own codes in Spanish given context about the project.
• 251 causal links were found.
• We also auto-coded sentiment: blue arrowheads for positive contributions, red for negative.

Step 4: Answering the research questions

Using the filters in the app to build maps that answered:

• What was the immediate impact on respondents’ lives from gender discrimination?
• What is the causal network from gender discrimination?
• Which factors are mentioned most often?

We also used the AI Answers feature to dig further into the interviews. AI Answers works independently of causal coding.

What the researcher said

“The type of questions asked, ‘what causes what’, were equally linked to methodological innovation. The results portrayed how gender barriers are intertwined in domains ranging from higher STEM education to the performance of new professionals and technicians once they enter the labour market, reaching deeper explanations and social impact.”

Javiera Cienfuegos, Senior Researcher