Skip to main content

What Voxaa brings you, in four points.

Conversational surveys on WhatsApp, in writing or by voice, in your respondents’ language, with results you can see from day one.

Results before fieldwork ends

Stop waiting for the end of fieldwork to see your results.

Response rate, completion, results per question, filters by city, neighbourhood and age bracket, comparison by test cell. You can fix a misunderstood question or top up a quota on day one, not three weeks later.

  • Soft launch 5–10%
  • Quotas by city
  • Quality report per campaign

Campaign · 20 g coffee sachet · Cells A & B

Updated just now

Q6 · Purchase intent (top 2 box), by price tested

  • Low price79%
  • Mid price58%
  • High price26%

Themes detected in verbatims

  • Handy size for corner shops · 41%
  • Affordable price · 27%
  • Taste too weak · 14%
Voices and languages

Interview out loud, in their language, the respondents text cannot reach.

Where reading is a barrier, voice opens the door: voices recruited and recorded in the local language ask your questions out loud, and respondents answer out loud too. It is access to the respondents text questionnaires cannot reach, where others do not go.

  • French
  • Wolof
  • Lingala
  • Bambara
  • Pidgin
  • Dialectal Arabic
Scripted or AI

Scripted for reliability. AI when you want to know why.

Pick a mode below: you see exactly what the respondent receives, and what you pay.

Mode 1 · included in every planLowest cost per interview

Your questionnaire, word for word, 1,000 times over

The engine runs your questions exactly as you wrote them: no improvisation, no call to a language model at every message. Predictable behaviour, near-zero running cost, results comparable from one wave to the next.

Behaviour
Deterministic: same question, same order, same validation for everyone
Best for
U&A, trackers, NPS, long questionnaires and repeated waves
Cost
WhatsApp message credits only
Output
Clean data, directly comparable across waves

Skip logic, validation by question type, a nudge on an unreadable answer, drop-out handling, full conversation log.

Flow preview · scripted mode

Q7

How often do you buy sachet coffee?

Single choice · 4 options · strict validation

Valid answer?

Yes → Q8No → scheduled nudge, back to Q7
Q8

Which brand do you buy most often?

If "never" at Q7 → skip to Q12

Q14 · End of flow — interview saved, quality score computed.

Mode 2 · premium optionBilled per probe, enabled question by question

The "why?" question, asked like a real interviewer

On open questions only, the agent probes, rephrases and digs deeper in natural language, one or two probes that you configure. You get usable verbatims instead of "it's fine".

Behaviour
Adaptive probing, only on the questions you enable
Best for
Concept tests, packaging, ad pre-tests, large-scale qualitative studies
Cost
WhatsApp messages + AI credits, capped per campaign
Output
Rich verbatims, themes detected automatically

Strict guardrails: the agent never leaves the scope of the survey, gives no opinion and collects nothing that is not in your questionnaire.

Conversation preview · AI mode

What did you think of this new packaging?

It's fine.

Thanks! What exactly is fine: the colour, the size, the price shown?

AI probe · 1/2

The size mostly, it fits in a bag and the corner shop can sell it one at a time.

Got it. And what would make you hesitate to buy it?

Follow-up · 2/2

If the price goes up too much, I go back to the old sachet.

Conversation or form

Two display modes, one channel.

Whatever the survey mode, you choose how each question appears on the respondent's phone.

  • Conversation mode

    One question at a time, like a chat with a research lead.

    Best for diary studies, qualitative follow-up and audiences unused to surveys.

    • Automatic probing on open answers

    • Voice notes accepted on any question

    • Highest completion rate

  • Form mode

    Several questions per screen, right inside WhatsApp, without leaving the app.

    Best for large-scale quantitative work, screeners and big panels.

    • Native WhatsApp interface

    • Closed questions, scales and grids

    • Same brief, from a quick test to a national panel