Organized by what you're trying to do — writing a brief, picking personas, setting up an interview — not a feature list.
Complete demonstration of how users conduct business research with atypica.AI through a real-world scenario
atypica digs into why people think what they think, with AI personas as interviewees — qualitative research in hours, not weeks. Describe your goal, approve the plan, AI runs it.
Click Start New Study, describe your goal in plain language, answer a few follow-ups from Plan Mode, confirm the draft plan. Setup takes 5–10 minutes; the research runs in the background.
Brief length barely affects the result — Plan Mode fills in missing detail through conversation. A clear one-sentence brief produces nearly the same plan as a detailed one, just with more clarifying rounds.
Read the executive summary first, then the recommendations. Only go deeper into full transcripts when you need to convince someone or explain a surprising finding. Most decisions need just the first two pages.
Not by default. Reports are written narratives — findings, recommendations, persona cards, transcripts. The data is qualitative, not statistical, so traditional charts don't fit.
Every report is available as an online view, a PDF download, or a public share link that opens without login. Choose based on audience and sensitivity.
A Panel is a saved group of AI personas you build once and reuse. Create it, then start research projects inside it whenever you have a question. Three steps from setup to results.
Describe the audience, AI runs 5–10 rounds of public-post scanning over 1–2 days, and you get a structured profile plus custom personas ready to interview.
Two separate sources: social media scanning pulls real public posts from Xiaohongshu, Douyin, X, Instagram, and TikTok; Interview and Discussion use the pre-built persona library, independent of live social data.
Nothing to configure. AI picks up your industry, role, and past research from conversations and remembers them across sessions. View or edit anytime under Settings.
Describe who you need in plain language and the library returns a ranked list. Pick 5–10 that span different concerns, not near-duplicates of the same person.
For most studies, use the public library — pre-built and ready in seconds. Build custom personas only when you need to simulate real customers, a niche audience, or private data.
Two main paths: upload an interview transcript, or run a Scout task. Upload if you have transcripts; Scout builds personas from observed social behavior if you don't.
PDF format, drag into the persona creation page, AI extracts in 1–3 minutes. Quality scales with transcript depth: 5,000+ words is rich, under 1,000 is thin.
Not in one click — upload one at a time, 1–3 minutes each. For many at once, running a Scout task and building from observed behavior is the cleaner alternative.
Your library holds bookmarked public personas and ones you built from interviews or Scout. Without upkeep it becomes a graveyard of stale profiles that drags down every project using them.
The actual click flow for running your first AI Interview — from creating the project to reading the report. Five steps, about 10 minutes.
5 is the standard starting point. The system supports up to 10 per interview — more isn't always better.
Yes — that's Discussion Chat. Default Interview Chat keeps each persona's conversation independent; Discussion Chat has personas talk with each other, guided by an AI moderator.
A Discussion Chat project runs 3–8 personas through a guided conversation. Who you invite and which format you pick change the output more than you'd expect.
Both produce insight, but answer different questions. Interview goes deep on why; Research goes wide on what's happening in the market.
Once an interview finishes, the platform generates a report. This guide walks through what you'll see and where to find each piece.
Sage isn't for interviewing AI personas — that's the Interview feature. Sage builds an AI expert from your materials so you can ask it questions anytime.
Asking ChatGPT to play a persona looks similar on the surface. The difference is what happens after the first conversation — whether the expert improves or stays frozen.
Complete demonstration of how users conduct business research with atypica.AI through a real-world scenario
Free lets you try every core feature with monthly limits. Paid removes the caps and adds team collaboration — the decision is mostly about volume and whether you work alone.
Your account settings has everything: usage, plan, and where to manage subscriptions or download invoices.
Inviting colleagues requires a Team plan. Free and Professional plans are single-user by design.
The Team plan lets members share work and build up a library over time. What's shareable depends on the feature — this guide covers what's currently shared and what isn't.
You upload sensitive transcripts, persona definitions, and research questions. This guide covers what's done with your data, what isn't, and what control you have.
The platform generates reports but has no built-in editor. Export, edit with your own tool, ship the final version. This guide covers export formats and editing patterns.
Closing the browser doesn't stop your research. The platform runs work in the background and notifies you when there's something to see.