Welcome Message & Consent

The welcome message is the first thing a participant reads. With consent enabled, they must agree to it before the interview starts.

September 26, 2026

Where it is

Welcome in the sidebar (/survey-welcome).

The page

  • Welcome Message — the text participants see. A new study starts with a short default ("Welcome to the survey. Your participation is voluntary…").
  • Generate AI Suggestion — drafts a message and consent statement from your study.
  • Require Consent — when on, the participant must answer Yes before any question is asked. Declining ends the session.
  • Translations — pick a language, add it, then translate and save.
  • Save Changes — stores the message and the consent setting.

What participants see

The message opens the chat, followed by a Yes/No prompt when consent is required:

Hi! Thanks for helping us understand meal-kit subscriptions. This chat takes
about 10 minutes. Your participation is voluntary and you can stop at any time.
Your answers are used only for research.

Please reply with Yes or No (or the number):
  1. Yes
  2. No

Consent is recorded as its own turn in the transcript (consent_selection).

Writing a good one

Keep it to a short paragraph or two and cover:

  • who is running the study and how to reach them
  • what it is about
  • how long it takes
  • that participation is voluntary and can stop at any time
  • what happens to the answers
  • compensation, if any

If you enabled Session Recording on the Details page, say so here. The participant sees a recording notice, but your consent text still has to state it.

AI drafting

Generate AI Suggestion builds a message from the study type, title, description, research question, expected duration (from the question count), your researcher name and contact from your profile, and any benefits and risks recorded on the study.

Known issue (September 2026). This currently fails with "Failed to generate welcome message". Your existing text is left untouched when it fails, so nothing is lost — but until it is fixed, write the message yourself.

Translations

  1. Choose a language from the dropdown.
  2. Click Add Translation — a row appears with a text box and two buttons.
  3. Click the wand to translate your English message, or type your own.
  4. Click save on that row.

One translation per language. Participants get the translation matching the study language they are using.

With MCP

get_welcome   {"study_id": "…"}
set_welcome   {"study_id": "…", "message": "Hi! Thanks for helping us…", "consent": true}
list_welcome_translations  {"study_id": "…"}
upsert_welcome_translation {"study_id": "…", "target_language": "Spanish",
                            "translation_text": "¡Hola! Gracias por…"}
delete_welcome_translation {"study_id": "…", "translation_id": "…"}

set_welcome replaces the message. upsert_welcome_translation stores one translation per target language — calling it again for the same language overwrites.

There is no MCP tool for AI drafting or machine translation; an agent should write the text (in each language) and pass it in.

Compliance

Deutero gives you the mechanism; the wording is yours. If your work is covered by an ethics board, GDPR, HIPAA or similar, have the consent text approved and keep your own record. The platform does not store a separate signed consent artifact — the agreement is the recorded turn in the transcript.


Last updated: September 2026