Free during early access

AI interviews, in place of forms

Bring the question. AI brings the craft.

The handbook, every interview, every close reading — AI carries the whole craft of research.

When to read, whom, and which version to keep — the judgment stays yours.

One study runs through this page: one counselor, one question, 24 students, three years.

01The verdict, up front

Surveys are fundamentally broken.

Everyday research comes in two forms: send a form, or hire interviewers. A form locks its questions before you know what to ask — miss one angle and the data is dead. Real interviews catch the truth, at $50–200 a session — never affordable for everyday research. This isn't about better tools. The form itself is broken.

Real interviews · always existed

per interview

$50–200

the process

schedule · transcribe · code

one wave, 24 people

$1,200–4,800

Not the interviewer's fault — they just never fit everyday research.

The form · the compromise

Q7 · How satisfied are you with your program?

Very satisfied
Somewhat satisfied
Not really

Q8 · Anything else? (optional)

That tick was supposed to hold this sentence.

“Some days I honestly love this program. But the workload nearly moved me out last year.”

Design is guesswork

old wayQuestions locked before you start; miss a crucial angle and the data is useless.

new wayOne sentence, co-authored into a full handbook with the Design AI.

Filling is tedious

old wayMandatory word counts, irrelevant follow-ups; respondents click randomly to finish.

new wayA real conversation — tap when hurried, go deep when engaged.

Analysis is shallow

old wayMechanical percentages; the why gets buried.

new wayA deep reading per person, then a synthesis across them.

Expected answers only confirm what you already suspected. The valuable part is what respondents bring up on their own — your blind spots.

The full research lifecycle, reimagined.

02Craft nº1 · design

Not from a blank page.

Her entire brief

Her

I want to know how this year has treated our students.

Design AI

The flip side of that question is the moment of almost leaving. I'll write the interviewer as someone who waits — never pushing — and start the outline from each student's concrete days. Here's a first draft; you edit.

Three exchanges later, the handbook looked like this — she reworded two lines.

Interview handbook · draft 2 · AI-drafted, edited twice by her

Interviewer persona

Warm, never pushing. When a student pauses, wait.

Theme outline

  1. 1How the year went — start from one concrete day
  2. 2What weighs most — coursework, relationships, money
  3. 3The moment of almost leaving — and what came after

Other tools help you collect and display. None helps you design the research itself.

03Craft nº2 · interview

Send one link.24real interviews.

The AI interviewer talks with each student one by one — chasing surprises, hearing what goes unsaid.

no progress bar · no endless checkboxes · a few dollars a session · any language

Transcript · wave 1 · 2026 · anonymized ✓

R-0427

Interviewer

You ticked “somewhat satisfied.” What would it take to reach “very”?

Respondent

Honestly, the coursework — last year it nearly moved me out of the program.

flagged: coursework → carries into wave 2 next year

Interviewer

“Nearly moved out” — what kept you?

Respondent

I decided to hold on for one more semester. Then it slowly got better.

tension: staying vs. leaving → into the reading

Multi-model anonymization · every session · the product's only automatic step

04Craft nº3 · the reading

Not a dashboard. Documents.

AI-written · on her trigger · cites 3 turns

Portrait

Second-year, engineering. Short answers, always concrete; the words grew when “staying” came up.

Tension

Staying, or leaving. The thought of switching was real; so was the decision to hold on one more semester. Both are his.

Wants

Not a lighter load — a visible end to it.

AI synthesis · reads the 24 readings · she selected all

7

nearly left

6

stayed — not for relief

Coursework is the heaviest theme (19/24). But among the 7 who nearly left, 6 stayed not for relief — but for a person, or a course.

A finding no form could ask for; it exists only where someone reads it out.

A re-run is a new version, never an overwrite.

The craft is AI's. The judgment is yours.

Analysis runs only when you start it. When to read, whom, and what to keep — your call.

Your data never trains anything

How privacy is structural →

05Next year

The same interviewer, who remembers everyone.

Next year, send the link again. The interviewer arrives carrying what each person said last time, and picks up where it left off — one continuous, years-long conversation. Snapshots go stale; a trajectory compounds. It is also why institutions stay.

Same respondent · three annual waves

R-0427

Wave 1 · 2026

Interviewer

How was the year, really?

Respondent

Hard. The coursework nearly moved me out.

Wave 2 · 2027

Interviewer

Last year you said the workload nearly moved you out — how does that choice look now?

Respondent

I stayed, and I'm glad. It eased after second year.

Wave 3 · 2028

Interviewer

Two years on — does switching still come up?

Respondent

Never again. This path became mine.

Other tools: a snapshot, every time.

Qualiask: a trajectory, over years.

Talk, don't tick.

Bring the question. AI brings the craft.

Create a project, say your question, send the link. Understand your users, track a cohort across years, or run real research where no team exists — free during beta.