Getting started

Last updated 25 Oct 2026 · by Surge Software

1. Create a study

Go to Studies → New study and describe your research goal in one or two sentences.
Pick interview mode (voice, video, text), languages, incentive and target completes.
Add image or video stimuli to run a concept or ad test — each gets its own section in the guide.
Add a prototype (a Figma prototype link, or any URL starting with prototype:) for a usability task: participants use it inside the interview, think aloud, and rate ease 1–5. A failed task triggers an AI follow-up.

The AI drafts sections, questions, probes and a time budget. Edit any question and its type (open, rating, NPS, choice).
Follow-up budget per question controls how deep the AI probes (0–4).
Every save creates a new guide version; restore any earlier version from the Guide tab.
Screener rules disqualify people who don't match; screen-outs are free. Quotas cap each answer (e.g. max 4 from Metro cities) or age band (18-30=5). The Guide tab shows live quota fill.

Open Pilot to watch an AI persona take the interview by voice. Fix confusing questions before paying for real people.
Synthetic interviews are labelled and never counted in billing.

Share the participant link (email, chat, in-app) or invite people you add by email from the Participants tab.
Screener quotas and screen-outs work today.
Panel recruiting (Prolific, Cint, Surge India Panel, B2B) is coming soon. Panel orders today run in sandbox mode so you can test the whole flow, including completion and screen-out return codes.
Participants give their UPI ID on the thank-you screen. Approve incentives, then download the payout CSV and pay them yourself.

Voice interviews keep each spoken answer; video interviews keep a camera clip per answer. Clips play on the transcript and in the highlight reel.
The app, database and clips are hosted on our own servers in Bengaluru.
Settings → Data retention deletes clips older than N days. Participants → Erase removes one person's transcript, clips and personal data (DPDP right to erasure).

Click Analyse to build themes, highlights and findings. Every finding cites the transcript turns it came from; citations are verified word-for-word.
Fraud scores (0–100) explain each signal. Interviews over your threshold (default 70) are excluded and not billed.

Export the report as Markdown/HTML, play the highlight reel, push findings to Slack, Notion or Jira.
Turn findings into decisions and track them to a measured outcome.

Settings → API & MCP creates a workspace key. REST: GET /api/v1/studies, /api/v1/studies/{id}/findings, /api/v1/studies/{id}/report, /api/v1/search?q=, /api/v1/decisions with Authorization: Bearer si_….
MCP: add the server URL /api/mcp with the same Bearer header to Claude, ChatGPT or Cursor. Tools: list_studies, get_findings, search_interviews, get_report, list_decisions.
SSO: Settings → Single sign-on takes your OIDC issuer, client ID and secret. Register the redirect URI shown there with Okta, Entra ID or Google Workspace.

Participants you add by email get one invite and, if they have not started after 24 hours, one reminder. Every invite has an unsubscribe link.
Workspace owners get a "results ready" email when an analysis finishes. Password resets, welcome mail and at-cost payment receipts are sent automatically.
Notification mail is never sent between 02:00 and 06:00 IST.

AI runs on Surge's own servers first; Claude (outside India) is used only as a fallback. Email is sent through Resend (outside India).
Indian-language voice interviews are coming soon.
Participants give recording and data-use consent (DPDP Act 2023) before the interview starts.