Running moderated and AI-moderated interview studies

Moderated and AI-moderated interviews add a qualitative layer to your research, helping you uncover the context, motivations, and expectations behind what people do in your product. In Maze, interview studies give you flexible ways to run and analyze conversations at scale—whether you’re moderating or working with an AI copilot.

In this chapter, we'll cover:

Moderated interview studies are live research sessions where you and a participant talk in real time, usually over video or audio, using a flexible discussion guide. You might show a prototype, concept, or idea in Maze, then ask follow-up questions to understand what people think, feel, and do as they move through it.

This enables:

With a human moderator in the session, you can clarify unclear answers, ask follow-up questions when someone hesitates or seems confused, and adapt your script if a task is harder or easier than expected.

This makes moderated user interviews especially useful when you’re exploring new problems, testing early ideas, or trying to understand specific behaviors and emotions.

Traditionally, moderated research required juggling separate tools for scheduling, video calls, recording, and transcription. Maze Interview Studies—Maze’s moderated interview solution—offers an end‑to‑end workflow that brings scheduling, conferencing, recording, transcription, and analysis into one place.

Maze supports the full moderated interview lifecycle, including:

With everything centralized in Maze, moderated research becomes easier to scale and share across teams.

AI-moderated interview studies use an AI moderator to lead or support live research conversations with participants, following a discussion guide you define. The AI moderator prompts participants, reacts to their answers, and keeps the conversation focused on your pre-defined goals.

AI-moderated interview studies are best suited to early-stage, generative research at scale, such as exploring a new market, understanding how people currently solve a problem, or testing how relevant a potential problem is for a specific audience.

Maze’s AI moderator runs interviews for you, so you can get rich qualitative insight without moderating every session yourself.

AI moderator helps you:

Both moderated and AI-moderated interviews help you learn from real conversations with participants, but they are better suited to different types of research goals.

Setting up a moderated interview in Maze is fast and easy. Participant selection, scheduling, and reminders all live in a single workflow, so you can move quickly from setup to live conversations.

Start by creating a new Moderated interview. Give the interview study a clear, descriptive name, and confirm key details like the expected session length and language so everyone on your team understands what this study is for.

Next, configure how participants will be invited and booked in.

In the recruitment settings, you connect your calendar or define specific time windows so Maze can show only the slots you want to offer. You can set limits for the total number of sessions and how many you want to run per day, and add a short screener if you need to qualify people before they book—especially useful when you’re recruiting via a panel.

When this is ready, Maze gives you a recruitment link that you can share through your own channels, in-product prompts, or panel providers.

You then decide how you’ll meet participants. Maze lets you use its built-in video conferencing or connect external tools like Zoom, Google Meet, or Microsoft Teams, and it automatically includes the right meeting link in each participant’s invitation.

At this point, you can also choose whether to record sessions. Turning on automatic recording is recommended if you want transcripts and highlights generated for you after the call.

As people start to book, upcoming interviews appear in your study’s sessions (the individual scheduled interview calls for that study) or the Recruit view. You join each call from Maze, greet the participant, and follow your discussion guide, while observers can quietly watch without being on stage with the participant.

And if you’ve enabled recording, you don’t need to worry about detailed note-taking during the conversation—Maze captures the session for you to review later.

Once sessions are complete, you’ll find recordings and transcripts in the study’s Session area. From there, you can review each conversation, clean up any transcript errors that matter, and save key moments for later analysis. Maze AI automatically generates transcripts, summaries, and highlights, reducing the manual work after each interview.

Creating an AI-moderated study follows a guided, step-by-step flow designed to help you define strong research goals, even if you’re starting from scratch.

From your project dashboard, select Create new study → AI-moderated. You’ll be prompted to choose a structure:

You can also adjust language preferences by setting participation language (what language participants speak) and analysis language (what language insights are synthesized in).

Next, you’ll define your research focus by:

This input is used to generate your initial research goals.

Context helps the AI moderator ask smarter, more relevant follow-up questions. You can add:

The more relevant context you provide, the more tailored the conversation guide will be.

Maze generates a set of research goals based on your inputs. These goals become the backbone of the AI-led conversation. You can:

Each goal becomes a section in the interview plan, with time estimates to help manage study length.

Once your goals are set, Maze creates a structured interview plan that includes:

You can preview the study to experience the AI moderator exactly as participants will—listening to how questions are asked and how the conversation flows.

After each moderated or AI-moderated interview, Maze helps teams review what happened, extract evidence, and package the findings in a format stakeholders can understand and use. Across both study types, Maze AI can generate transcripts, summaries, and suggested highlights, which gives teams a faster starting point for analysis.

From there, researchers can review sessions, group highlights into themes, and build a clearer picture of what participants said across interviews.

In moderated studies, your team usually takes a more hands-on role in deciding which highlights to keep, how to name themes, and how to shape the story.

In AI-moderated studies, Maze can identify theme clusters and sentiment across sessions, giving you more structure upfront. Either way, you review the output, edit, and refine it before sharing it widely.

Once findings are organized, teams can use Maze’s automatically generated reports to present them to stakeholders. These reports include an executive summary, participant-level views, and theme slides built from highlights and key findings. Researchers can edit summaries, refine findings, adjust highlights, and hide themes they do not want to include in the final report.

Maze also lets teams share findings at different levels. You can share the full report, a specific report slide, or individual pieces of evidence such as themes, highlights, sessions, and highlight reels.

To share a full report or a specific slide, open the report and click Share report to copy the share link. You can also copy an embed link to view the report in another tool. Report links can be private, so only authenticated team members can access them, or public, so anyone with the link can view them. Highlight reels can also be shared by link or embed, or downloaded as video files.

That makes reporting much easier for teams that need to socialize insights across product, design, and leadership.

Now that you’ve set up and run your studies, the next step is understanding what the results are telling you. In the next chapter, we look at how to read Maze reports—across metrics, open-ended responses, themes, and highlight clips—so you can turn raw data into actionable insights for your team.