Product
AI interview platform for qualitative research
Fieldwork Interviews gives research teams a way to run adaptive, in-depth qualitative interviews at scale—without moderator scheduling, availability constraints, or consistency problems. Sofi conducts every session using your study design, probes weak answers, and returns structured outputs so analysis starts immediately.
How does Sofi conduct an interview?
Sofi follows the interview structure you define: the topics to cover, the interview depth setting for each topic, and the rules for when to move between topics. She is not improvising a new methodology each session—she is executing the study you approved.
During the conversation, Sofi maintains a live understanding of what she has learned about each topic. That lets her calibrate depth: she spends more time where answers are thin or contradictory, and she moves on when a topic meets what you defined as "enough" before advancing.
When a participant gives a short or evasive answer, Sofi is designed to notice it and probe directly—without filler praise, without accepting vagueness, and without stacking two questions in one turn. Her register stays professional and efficient, which respects participant time and keeps the transcript usable for analysis.
Sofi also adapts her word budget to how people communicate: more concise participants get tighter questions; more verbose participants get room to elaborate without losing structure. The goal is depth where it matters, not uniform talk time for its own sake.
What do researchers control?
You start from a research brief. Sofi generates a proposed study structure—topics, depth settings, and interview logic—which you review and edit before launch. Nothing goes live without researcher approval.
You set screening questions to qualify participants before the main interview begins, and you can set completion redirect URLs when you are routing participants from a panel or recruitment partner.
Studies can be Draft (closed to participants) or Live (open). You can pause, archive, or adjust studies as your programme evolves, without losing a clean record of what was fielded when.
What comes back after sessions complete?
Full transcripts appear as sessions complete. Each session also receives an automated quality check—a structured score that helps you spot interviews that may need review without reading every line first.
The topic coverage tracker shows which topics were resolved, partially covered, or missed for each participant. At the study level, gap reporting highlights themes that consistently stayed underexplored, so you can adjust the study design or recruitment rather than discovering the blind spot in a deck three weeks later.
Exports are available as structured data; CSV export for transcripts and related fields is available on Growth plans and above. Automatic analysis after each interview runs in the background, so synthesis can begin while fieldwork is still in motion.
What does a UX study look like in practice?
A UX researcher at a fintech company is testing a redesigned onboarding flow for new business account holders. The team needs to know where confidence drops and what is driving abandonment at document upload—not just that drop-off exists in analytics.
The brief
"We have redesigned onboarding for new business account holders. We need to understand where users lose confidence in the process and what is driving drop-off at the document upload step."
What Sofi generates
- First impressions — quick check-in on initial reactions
- Document upload experience — in-depth behavioural exploration
- Confidence and trust signals — problem and solution framing
- What would have helped — concept-style validation of improvements
Example conversation · not live data
Walk me through what you were thinking when you got to the document upload step.
It was fine. I just uploaded and moved on.
That is quite short. What felt easy about it, and what—if anything—made you hesitate before you hit submit?
Sofi · probing a brief answer
I was not sure whether my scan would be rejected. The wording sounded like any mistake would lock the account.
What comes back
Fourteen transcripts with the document upload topic resolved in eleven sessions; three sessions flagged for review. Emerging themes include unclear instructions, fear of rejection, and brand trust—ready for a design critique the same week fieldwork started.
Which interview styles does Fieldwork support?
Exploratory / behavioural interviews are for when you need depth on motivations, habits, and trade-offs. Use them for discovery when you cannot yet name the right solution space.
Quick pulse checks are short, focused sessions when you need a fast read on a narrow question without running a full behavioural arc.
Concept and hypothesis testing helps you validate whether an idea lands, where confusion appears, and what would need to change before you commit design or engineering capacity.
Journey and narrative research captures stories over time: sequences of events, hand-offs, and emotional beats that quantitative funnels flatten away.
Problem and solution exploration structures how people frame problems, what they have already tried, and what would count as a meaningful fix—ideal for positioning and roadmap debates.
Frequently asked questions
What types of qualitative research can Fieldwork Interviews support?
Fieldwork supports exploratory discovery interviews, concept and hypothesis tests, customer journey mapping, quick pulse checks, and in-depth behavioural research. Each study type uses a different interview structure—you select the type when creating a study and Sofi adapts her approach accordingly. Teams also build custom study structures for research that does not fit a standard template. That flexibility matters when your brief spans multiple decision types in one programme.
How is this different from a survey or questionnaire?
Surveys send fixed questions in a fixed order and typically accept whatever response comes back without follow-up. Sofi conducts a real conversation: she follows up on interesting answers, probes vague ones, and redirects if a participant goes off track. The output is a transcript and structured analysis grounded in what people actually said, not only a response distribution. That difference is what makes the method qualitative rather than quantitative.
Can we control what Sofi asks and how deeply she probes?
Yes. You define the interview topics, the interview depth setting for each topic—from a quick check-in to a full behavioural deep dive—and the rules for when Sofi should move between topics. You also set guardrails for areas Sofi should not enter. The interview logic is yours; Sofi executes it consistently across every session. Researchers review and edit the structure before anything goes live.
How long does a typical interview take?
Interview length depends on the study design and the participant. A quick pulse study with two or three topics typically runs about eight to twelve minutes. A full behavioural interview with five or six topics often runs twenty-five to forty minutes. Sofi calibrates depth to the participant communication style, so more verbose participants tend to produce longer, richer sessions while brief communicators still get appropriate follow-up without padding.
What happens after an interview session is completed?
As soon as a session ends, Fieldwork runs automatic background processing: generating a transcript, running an automated quality check on the interview, updating topic coverage, and flagging any topics that were not fully resolved. Exports are available without a separate cleanup step. You do not need to wait for a batch export job before starting synthesis.
Is there a free tier?
Yes. The Starter plan includes ten completed interviews per month at no cost, with no credit card required. That is enough to run a real study and evaluate whether Fieldwork fits your workflow before moving to a paid plan. Billing is based on completed sessions, not abandoned or very short sessions.
Related: UX research at scale, ResearchOps workflows, In-product interview SDK, and pricing on the home page.