AIMI whitepaper: a new methodology for qualitative research at scale

What are AI-Moderated Interviews? How they work, and why they matter (whitepaper)

In 2024, Glaut coined the term AIMI - AI-Moderated Interviews - to introduce a new research methodology made possible by a new generation of natural language technology.

Just as CATI (Computer-Assisted Telephone Interviewing) and CAWI (Computer-Assisted Web Interviewing) marked previous leaps in data collection, AIMI mark a new one: open-ended, AI-moderated, one-on-one interviews that adapt to each participant in real time and scale, without sacrificing depth.

We’ve now published our first whitepaper: “AI-Moderated Interviews in Practice: Early Evidence and Open Questions”

What you’ll learn inside

This AIMI whitepaper is a practitioner’s guide: clear, evidence-based, and built for researchers.

• What AIMIs are: a distinct, flexible methodology that merges qualitative depth with quantitative scale
• How they work: modular design, real-time adaptability, safeguards, and built-in interpretative layers
• Full results from our first comparative study with surveys, including:
  • +129% words per answer
  • -53.6% gibberish reduction
  • +56% valid completions
• Case examples: kids ages 3–13 completing voice interviews solo and a 95% drop in coding time for a major e-commerce brand.

Download the AIMI whitepaper

A new methodology not only a new tool

AIMIs (AI-Moderated interviews) aren’t “surveys with a friendlier interface.” They rethink how conversations are designed, guided, and analyzed without a human moderator in the room, but with the researcher fully in control. Every response is traceable, every theme is editable, and every insight stays accountable to the original voice that shared it.

What’s next?

We don’t see AIMI as a replacement for in-depth interviews or a perfect fit for every case, but we do see them opening up a new methodological middle ground. This paper shares the early signals, the open questions, and where we’re heading next from format comparisons to interpretative benchmarking.