There was significant improvement in the responses when using the AI chat bot to instigate more comprehensive responses from the respondent. Additionally, there were also advantages to limiting the number of probes provided by the AI chat bot. The use of the AI chat bot as part of the survey design resulted in much more powerful data to fuel business decision-making than the control method of a traditional open text response, even with a quality meter score attached to the text box.
The respondents who were provided with the AI chat bot as part of their survey produced better feedback than traditional open-ended questions. The chat bot with two additional probe questions produced 2.5x more feedback, and the one with four probe questions produced 5x more feedback.
The responses from each of the four additional probe questions were compared using the Flesch Kincaid Grade Level formula and resulted in elevated grade levels by at least one full grade level versus traditional open-ended methods. This indicates a more detailed and thoughtful response to the question asked in the survey.
When asked, the survey participants who responded using the AI chat bot said they were able to share everything they wanted to share (62% vs 44% agree) and that they found it easy to share ideas this way (61% vs 49% agree).
Using AI tools as part of a research design can dramatically improve the quality and depth of the research insights compared to traditional methods. As AI in research is a recommendation to consider not every project and not every open-end requires nuanced data that conversational probing provides.