Chatbot-Administered Surveys in Marketing: A Literature Review

Abstract

This article reviews investigations of online survey chatbots across disciplines and offers insights into their potential to encourage innovative and practical approaches to marketing research surveys. Online surveys deployed in service practice can effectively capture customer feedback. Using chatbots to administer such surveys can have desirable outcomes, although specific advantages and disadvantages are not clear at this point. Based on a meta-analysis of prior research on the use of chatbots in online surveys, this article presents a typology of relevant themes and avenues for further research. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method initially yielded 1,827 unique papers from Scopus, of which 30 qualify for in-depth analysis. The analysis identifies five key survey administration capabilities of chatbots: data collection, conversational interfaces, interactivity, humanization, and effective response generation. It also highlights the challenges and research gaps associated with each capability.

Keywords

chatbot online survey marketing research literature review prisma thematic content analysis

Introduction

Market transformations that are driven by artificial intelligence (AI) include the expanded use of chatbots, which are computer programs that replicate human communication in text- or voice-based dialogue systems (Gaikwad et al., 2018). The global chatbot market continues to grow, with an expected value of US$27.29 billion by 2030 (Tyumeneva, 2026). Chatbots can handle and promptly respond to simple queries in call centers, while human employees handle more complex issues, thus promising substantial reductions in labor costs. It is reported that Chatbots can respond quickly and answer an estimated 80% of frequently asked questions (Downie & Hayes, 2025).

Chatbots are also promising for online surveys (Jin et al., 2024) and may replace other, less efficient survey methods, such as in-person or telephone surveys. Chatbot surveys are continuously available (Araujo, 2020). In this sense, chatbots offer enhanced convenience (Kim et al., 2019) and efficiency (Belhaj et al., 2021). Current implementation of chatbots can engage in natural, conversational interactions with survey participants (Celino & Re Calegari, 2020), as exemplified by surveys embedded in Facebook Messenger, which encourage participation by appearing in a familiar setting for many consumers (Smutny & Schreiberova, 2020).

Extant research has examined the use of chatbots across various domains, including education (Belhaj et al., 2021), computer science (Jin et al., 2024), and health care (Te Pas et al., 2020). We conduct a literature review, using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) approach (Page et al., 2021). We review the application of chatbot surveys across these domains, focusing on applications relevant to market research. Marketing researchers can draw from established knowledge in other domains. By identifying 1,827 unique papers and further analyzing 30 highly relevant papers, we established five main themes and noted several related but less prominent themes.

Academically, the findings summarize the literature on the use of chatbots in market research surveys and outline an agenda for future research. While previous studies have examined chatbots from different perspectives, the literature review in this paper provides a more comprehensive overview of relevant themes than existing work. For instance, Dwivedi et al. (2021) did not investigate how chatbots can be implemented in specific contexts, such as online surveys, whereas Höhne et al. (2024) primarily examined the adoption of chatbots and their effects on outcomes (i.e., completion rates). Building on these studies, we contribute to the literature by providing a systematic synthesis of chatbot-administered surveys, offering structured insights into their design and implementation in online survey research. For marketers, this paper highlights the potential and limitations of chatbots as data collection tools and provides guidance on their implementation. The discussion section elaborates on the study’s implications for academic literature and marketing research.

The paper is structured as follows: first, we present the theoretical background; next, we describe how we applied PRISMA methodology to select papers, detailing how articles were identified, screened, deemed eligible, and included. Lastly, we apply thematic content analysis to evaluate the 30 key articles, identifying five main themes and five additional emerging themes. The analysis of primary themes underscores the benefits and limitations of chatbot capability dimensions and provides directions for both academic research and practice, as outlined in the discussion.

Theoretical Background: Chatbots

Contemporary chatbots can engage in humanlike dialogue and perform natural language processing (NLP). They can be integrated into various application programming interfaces (Tsai et al., 2021) to resolve customer requests and automate repetitive tasks, such as answering frequently asked questions or sending customer emails. When applied to online surveys, chatbots can also improve efficiency (Belhaj et al., 2021) by gathering vast amounts of data and streamlining the survey process, such as by explaining the purpose of the questionnaire and guiding participants as needed (Söderström et al., 2021). Furthermore, whereas conventional online surveys tend to use a formal language style, the casual language style often programmed into chatbots can resonate with participants’ expectations of social interactions, though it may also engender embarrassment (Kim et al., 2019).

Noting the many studies on the use of online chatbots, we also identify a gap: there is no clear overview of the current state of knowledge. Literature can provide insights into current knowledge of chatbot use in market research surveys, highlight promising avenues for future research, and offer guidelines for practitioners applying chatbots in market research. Based on this review, we address three focal research questions that drive our investigation:

Research question 1

What does current literature say about chatbots’ online survey capabilities?

Research question 2

What are the key benefits and limitations of chatbot implementation in online surveys, as identified by existing research?

Research question 3

What directions for research can be derived from the extant academic literature to guide ongoing chatbot implementation in online marketing surveys?

We undertake a systematic literature review to assess what is known, particularly the key advantages and limitations of chatbots, and to propose directions for further research.

PRISMA-Guided Selection of Relevant Papers

We applied the PRISMA framework, a structured and reproducible search strategy (Page et al., 2021), comprising identification, screening, eligibility, and inclusion. We searched across disciplines beyond marketing, including, for example computer science and information systems, for studies on chatbots in online surveys. By rigorously adhering to predefined criteria, we captured all relevant studies. Figure 1 depicts the step-by-step process for screening and excluding articles.

Figure 1.

Search results for PRISMA

In Step 1 (Identification), we first defined relevant research items and derived search keywords that are based on our research questions (Wohlin et al., 2020). Keywords and synonyms were selected to ensure a comprehensive search while minimizing irrelevant results. A preliminary search in Scopus using all keyword combinations in Table 1 retrieved 3,181 records. After removing 1,354 duplicates, 1,827 unique articles were retained for screening.

Table 1.

Keyword Combinations

Set 1	AND	Set 2	AND	Set 3
‘Chatbot’ OR ‘conversational agent’ OR ‘visual assistant’	AND	‘Survey’ OR ‘questionnaire’	AND	‘Respondent’ OR ‘customer’ OR ‘participant’ OR ‘patient'

In Step 2 (Screening), we applied filters to exclude topics that are not relevant for the use of chatbots in surveys, such as chatbots’ social presence in communication (Tsai et al., 2021). Journal articles were limited to 2013-2025, as we identified Hasler et al.’s (2013) as the earliest relevant study. Conference papers were limited to 2019-2025 to anticipate emerging research. Titles, abstracts, and reference lists were checked for reliability and credibility. Three screening exclusion criteria (SEC) were applied. Therefore, 603 articles remained for eligibility assessment.

SEC1: Articles that are non-English or outside the defined publication period, removing 99.

SEC2: Articles that do not focus on chatbots in online surveys, removing 1,093 whose abstracts, introductions, or focus fell outside our review scope (i.e., customer service, general AI systems).

SEC3: Articles that are not peer-reviewed, removing 32 articles.

In Step 3 (Eligibility), we further narrowed the selection based on journal credibility and empirical evidence, following Patino and Ferreira (2018) and Meline (2006). Two eligibility exclusion criteria (EEC) were applied, and 30 remained after this step.

EEC1: Articles that are not fully accessible (i.e., missing full texts), removing 384 articles.

EEC2: Journals that do not rank as A*, A, B, or C in the Australian Business Deans Council (ABDC) Journal Quality List, nor as Q1 (top 25%), Q2 (25-50%), or Q3 (50-75%) in the Scimago Journal & Country Rank, removing 189 articles.

Step 4 (Inclusion) ensured robustness by evaluating the 30 articles for relevance, focus on chatbots in online surveys, and methodological rigor. No articles were removed, ensuring our findings are based on high-quality, current, and thematically relevant sources. The Online Appendix summarizes these articles, including publication source (journal or conference), applied theories, research focus, survey approach, methodology, chatbot capabilities, and key findings.

The descriptive statistics reported in Figure 2 illustrate the publication year of the 30 articles, revealing increased research interest since 2019. Figure 3 shows an increased interest in the application of chatbots in surveys, in terms of the number of times each of the 30 selected articles is cited each year. Table 2 lists the number of cited articles by journal or conference, and Computers in Human Behavior publishes the largest number of papers on the application of chatbots in surveys.

Figure 2.

Publication year

Figure 3.

Total citations per year

Table 2.

Cited Articles by Journals/Conferences

Journal or conference	Number	Study
ACM Transactions on computer-human interaction	2	Xiao et al. (2020a); Zhou et al. (2019)
arXiv preprint	1	Tang et al. (2025)
Behavior research methods	1	Söderström et al. (2021)
CHI conference on human factors in computing systems proceedings	4	Kim et al. (2019); Xiao et al. (2020b); Jacobsen et al. (2025); Li et al. (2025)
Communication methods and measures	1	Zarouali et al. (2024)
Computation communication research	1	Araujo (2020)
Computers in human behavior	5	Hasler et al. (2013); Lucas et al. (2014); Pickard et al. (2016); Rhim et al. (2022); Ward and Pond (2015)
Frontiers in digital health	1	Soni et al. (2022)
Frontiers in psychology	1	Conrad et al. (2015)
IEEE International conference on data science and computer application	1	Guo et al. (2021)
Indonesian journal of electrical engineering and computer science	1	Belhaj et al. (2021)
International conference on human-computer interaction	1	Abbas et al. (2021)
International conference on intelligent user interfaces	1	Xiao et al. (2019)
International journal of human-computer studies	2	Celino and Re Calegari (2020); Jin et al. (2024)
International scientific conference on information technology and data related research	1	Duka and Njeguš (2021)
JMIR medical informatics	1	Te Pas et al. (2020)
Journal of development economics	1	Beam (2023)
Journal of information systems education	1	Gupta and Chen (2022)
Journal of medical internet research	1	Park et al. (2019)
Journal of service management	1	Sidaoui et al. (2020)
Science, education and innovations in the context of modern problems	1	Hassina and Hiba (2025)

Thematic Content Analysis

We conducted a thematic analysis of the 30 articles in Table 2 to identify themes and trends across studies. Using the method outlined by Cruzes and Dyba (2011), we began with an exploratory text analysis to identify common themes. This iterative process (Neuendorf, 2018) allowed us to refine themes iteratively and assess each study’s contribution, providing insights into the scope, depth, and breadth of the research. The first author analyzed all papers, while the other two authors derived themes from half of the papers each. Such qualitative evidence synthesis enhances the reliability of the findings.

From the 30 studies on chatbots in online surveys, we identified five key themes related to chatbots’ survey capabilities (Table 3). A key theme is mentioned in at least six articles. These key themes are data collection, conversational interface, interactivity, humanization, and effective responses (RQ1). For each theme, we examine its benefits and limitations (RQ2) and offer suggestions for research and development, including ideas for integrating chatbot technology into survey administration processes and addressing ethical implications, such as privacy protection (RQ3). We also found five less common but relevant themes: conversational styles, gamification, response delay, chatbot questionnaire contextuality, and comfortable information sharing. The Online Appendix provides a detailed summary for each article.

Table 3.

Five Key Themes and Related Studies

Capability theme	Definition	Study
Data collection	Embed a conversational chatbot into a survey to collect data	Abbas et al. (2021); Araujo (2020); Beam (2023); Celino and Re Calegari (2020); Hasler et al. (2013); Jin et al. (2024); Kim et al. (2019); Lucas et al. (2014); Park et al. (2019); Sidaoui et al. (2020); Te Pas et al. (2020); Xiao et al. (2019); Zarouali et al. (2024); Zhou et al. (2019); Jacobsen et al. (2025); Li et al. (2025); Tang et al. (2025); Hassina and Hiba (2025); Soni et al. (2022)
Conversational interface	A user survey interface where a chatbot simulates a human conversation to provide respondents with a dialogue-based experience	Belhaj et al. (2021); Rhim et al. (2022); Celino and Re Calegari (2020); Kim et al. (2019); Pickard et al. (2016); Xiao et al. (2020a, 2020b); Duka and Njeguš (2021); Jacobsen et al. (2025); Hassina and Hiba (2025); Tang et al. (2025)
Interactivity	Facilitate two-way interaction between respondents and a chatbot in the survey	Beam (2023); Belhaj et al. (2021); Celino and Re Calegari (2020); Guo et al. (2021); Gupta and Chen (2022); Jin et al. (2024); Kim et al. (2019); Lucas et al. (2014); Duka and Njeguš (2021); Conrad et al. (2015); Rhim et al. (2022); Soni et al. (2022); Ward and Pond (2015); Xiao et al. (2020a, 2020b); Jacobsen et al. (2025); Pickard et al. (2016); Li et al. (2025); Hassina and Hiba (2025); Tang et al. (2025); Zarouali et al. (2024); Araujo (2020)
Humanization	Exhibit humanlike qualities or the appearance of a chatbot in the survey	Conrad et al. (2015); Hasler et al. (2013); Kim et al. (2019); Rhim et al. (2022); Xiao et al. (2020a, 2020b); Zhou et al. (2019); Pickard et al. (2016); Tang et al. (2025); Celino and Re Calegari (2020); Zarouali et al. (2024)
Effective response	Receive high-quality, effective, and meaningful responses in the chatbot survey	Abbas et al. (2021); Beam (2023); Belhaj et al. (2021); Celino and Re Calegari (2020); Conrad et al. (2015); Guo et al. (2021); Hasler et al. (2013); Kim et al. (2019); Park et al. (2019); Söderström et al. (2021); Soni et al. (2022); Ward and Pond (2015); Xiao et al. (2020a, 2020b); Zhou et al. (2019); Jacobsen et al. (2025); Li et al. (2025); Tang et al. (2025); Sidaoui et al. (2020); Rhim et al. (2022); Jin et al. (2024)

Theme 1: Data Collection Capability

A growing body of literature examines the feasibility of using chatbots embedded in surveys to collect data, replacing traditional open- or closed-ended questions with conversational chatbot-based interview or questionnaire prompts (e.g., Kim et al., 2019; Xiao et al., 2020a; Zarouali et al., 2024). They show that chatbot-assisted data collection can enhance participant disclosure (Beam, 2023; Jin et al., 2024; Lucas et al., 2014; Te Pas et al., 2020; Zhou et al., 2019) and improve response quality compared with conventional online survey formats (Beam, 2023; Jin et al., 2024; Kim et al., 2019; Sidaoui et al., 2020). Positive outcomes are more frequently found when chatbots are used to administer structured or semi-structured questions. Under such conditions, chatbots are hypothesised to reduce satisficing behaviour and response biases by encouraging greater engagement and attentiveness (Beam, 2023; Hassina & Hiba, 2025).

These benefits are not consistently replicated across studies. Notably, using chatbots to collect large-scale survey data can be challenging, due to the inherent inaccuracies of NLP technologies. Therefore, chatbots may misinterpret respondents’ inputs or generate inappropriate follow-up responses, leading to dissatisfaction (Jacobsen et al., 2025; Kim et al., 2019; Li et al., 2025) and conversational disengagement (Hasler et al., 2013; Jin et al., 2024; Tang et al., 2025). These issues might ultimately diminish respondents’ willingness to answer subsequent questions, thereby compromising response quality, and challenge the claim that chatbots can reduce respondent satisficing behaviour.

This contradiction seems to be driven by differences in chatbot design (Jin et al., 2024) and interaction quality (Celino & Re Calegari, 2020; Xiao et al., 2020a). Some studies report that chatbots with clear conversational structures (Jin et al., 2024; Ward & Pond, 2015), limited or repeated probing (Li et al., 2025; Park et al., 2019; Zarouali et al., 2024), and constrained response formats (Jacobsen et al., 2025; Jin et al., 2024) function better and enhance interaction quality. This reduces cognitive load and minimizes misinterpretation. In contrast, when chatbots are required to interpret diverse open-ended responses or to dynamically adapt to respondents, interaction breakdowns become more likely (Park et al., 2019; Xiao et al., 2019). Respondents’ unfamiliarity with conversational survey formats (Lucas et al., 2014; Tang et al., 2025; Zarouali et al., 2024) can increase response times and this may result in shorter or less detailed answers (Zarouali et al., 2024). Similarly, if a chatbot continues to probe the same topic while ignoring cues to shift direction, it can generate frustration (Park et al., 2019), disrupt conversational flow, and ultimately lead to disengagement (Zarouali et al., 2024).

These findings suggest that the effectiveness of chatbots for data collection in online surveys depends heavily on their design and deployment. Chatbots seem more effective for structured data collection tasks, whereas their application to complex, open-ended survey contexts remain prone to interactional and technological limitations.

Based on this theme, we propose three research avenues. First, future research could explore the design of structured conversations in chatbot surveys to provide clear guidance and a sense of dialogue (Araujo, 2020; Jin et al., 2024; Ward & Pond, 2015). Researchers should draw attention to minimal or repeated probing in chatbot surveys because it can clarify questions or responses but may also cause fatigue when overused (Li et al., 2025; Park et al., 2019; Zarouali et al., 2024). Constrained responses may minimize ambiguity and misinterpretation (Jacobsen et al., 2025; Jin et al., 2024), whereas diverse open-ended responses are crucial for obtaining richer insights (Xiao et al., 2020a, 2020b; Zhou et al., 2019). Second, future studies could investigate how to predict and address interaction breakdowns when chatbots misunderstand inputs (Park et al., 2019; Xiao et al., 2019). Third, respondents’ unfamiliarity with conversational survey formats affects their expectations and subsequent behaviours (Lucas et al., 2014; Tang et al., 2025; Zarouali et al., 2024).

Theme 2: Conversational Interface Capability

By interacting via natural language in text-based or voice-based formats, chatbots are designed to simulate human dialogue and provide a more natural, engaging interaction experience for respondents (Celino & Re Calegari, 2020). A conversational survey aims to make participants feel as though they are engaging with a person rather than simply completing a traditional questionnaire (Duka & Njeguš, 2021; Kim et al., 2019). In such settings, chatbots offer a unique advantage by combining the efficiency of digital data collection with the interactive qualities of a human interviewer (Belhaj et al., 2021; Pickard et al., 2016). Chatbots facilitate quick, cost-effective collections of both numerical and textual data. These conversational interfaces create a natural and familiar medium for user expression, improve usability, and support flexible interactions that do not require a rigid path (Xiao et al., 2020a; 2020b). Using anthropomorphic and personalizing features can help capture user attention and build trust (Rhim et al., 2022; Tang et al., 2025).

If the chatbot effectively guides participants through a conversational survey, it provides clear and easy-to-understand information. Nevertheless, chatbots sometimes fail to address specific user demands or clarify survey features (Rhim et al., 2022). At times, a specific question of a respondent may necessitate intervention from a third-party actor, such as a human survey administrator (Belhaj et al., 2021). Therefore, although conversational interfaces are promoted as flexible and engaging, their actual operation often depends on pre-scripted conversational flows that cannot accommodate the unpredictability of real conversational behaviour (Jacobsen et al., 2025).

The benefits of conversational interfaces are often evaluated in controlled interactions where participants follow anticipated conversational paths. In contrast, studies that report limitations typically examine situations in which participants engage in conversation rather than in procedural tasks (Belhaj et al., 2021). In this case, chatbots cannot adapt to respondents, which undermines the expectations they create because of their different design from traditional surveys. Features that initially increase engagement might later undermine trust-building when expectations are not met (Rhim et al., 2022). Furthermore, requiring continuous internet connectivity leads to an additional challenge. Conversational interfaces aim to create an interactive experience for respondents, but a poor network connection can disrupt the flow of conversation between respondents and chatbots. Thus, the effectiveness of conversational interfaces is shaped by design and technological limitations.

Recent advances continue to leverage NLP to create generic conversational interfaces that support more humanlike interactions. Future research could investigate how respondents engage with chatbots to distinguish their behaviors between conversational and procedural modes (Belhaj et al., 2021) and whether the design of conversational interfaces meets respondents’ expectations for responsiveness and naturalness (Rhim et al., 2022). Furthermore, future studies could solve the practical challenge of internet connectivity by exploring how human-chatbot collaboration can be maintained effectively when connectivity issues occur.

Theme 3: Interactivity Capability

Interactive information exchanges are possible in a chatbot-enabled survey, such as repeated question elicitation. Furthermore, chatbots facilitate two-way communication, they can foster dialogue while still allowing users to maintain control over the flow and depth of the conversation. This contributes to reducing survey fatigue (Hassina & Hiba, 2025) and thereby enhances response quality (Beam, 2023; Guo et al., 2021; Jacobsen et al., 2025; Kim et al., 2019). Guo et al. (2021) find that participants remain positively engaged and willing to interact for extended periods if the chatbot demonstrates strong conversational capabilities. Chatbot interactivity enhances various other interaction outcomes, including user engagement (Belhaj et al., 2021; Celino & Re Calegari, 2020), favorable attitudes (Duka & Njeguš, 2021). Such benefits encourage positive affective, cognitive, and behavioral responses.

These findings assume that respondents view the chatbot as a conversational partner and are willing to engage; however, other literature highlights the negative effects on participant engagement in chatbot surveys. As defined by Dwyer et al. (1987), a disengaged state can occur at any point in a relationship and may lead to termination. In the context of survey chatbots, participants exhibit disengagement by expressing unfavorable attitudes toward the survey or the chatbot. This may undermine the system’s capabilities (Gupta & Chen, 2022; Li et al., 2025). For example, disengaged respondents might skip open-ended questions (Jacobsen et al., 2025; Zarouali et al., 2024), provide irrelevant or minimal answers (Xiao et al., 2020a; 2020b), or spend longer completion time than traditional surveys (Kim et al., 2019; Soni et al., 2022; Zarouali et al., 2024).

On another note, when respondents perceive the chatbot primarily as a survey tool rather than a conversational partner, interactivity may be perceived as time-consuming (Rhim et al., 2022; Zarouali et al., 2024) and cognitively demanding (Jacobsen et al., 2025), leading them to minimize participation or engagement. The effectiveness of interactivity depends on how respondents view the chatbot’s role in the process, which can either facilitate engagement or trigger disengagement.

In contemporary marketing research, personalized interactions and communication have become vital. Scholars could examine respondents’ perceptions of chatbots (as survey tools or conversational partners) and how these perceptions affect engagement and quality. Previous research indicates that respondents spend more time answering questions in chatbot surveys (Rhim et al., 2022; Zarouali et al., 2024), leading to survey fatigue or disengagement. Thus, future studies could adopt a balanced approach to interactivity and cognitive load to optimize respondents’ experience.

Theme 4: Humanization Capability

Relative to traditional online surveys, chatbots offer a more personalized approach that can enhance participants’ engagement (Conrad et al., 2015; Rhim et al., 2022; Xiao et al., 2020a, 2020b), willingness to respond, and response quality (Conrad et al., 2015; Xiao et al., 2020b). The perceived humanlike qualities of chatbots have encouraged respondents to share their thoughts and insights (Zarouali et al., 2024). In social media settings, chatbots provide interactive, personalized experiences by delivering interventions and can conduct surveys on these platforms (Kim et al., 2019; Rhim et al., 2022; Tang et al., 2025).

Anthropomorphic cues motivate greater self-disclosure among survey respondents, thereby eliciting open-ended, emotionally intense responses (Conrad et al., 2015; Hasler et al., 2013; Kim et al., 2019). Such increased self-disclosure likely stems from the chatbot’s ability to foster social bonding while preserving the respondent’s anonymity. While humanization capability encourages openness in low-sensitivity survey tasks, in sensitive survey contexts, respondents may become more cautious or even unwilling to disclose (Zhou et al., 2019).

Although a positive correlation exists between a chatbot’s verbal anthropomorphism and users’ affinity for it, this association holds only up to a point. As a chatbot’s interactions become excessively humanlike, users’ positive feelings diminish (Hasler et al., 2013; Kim et al., 2019; Zhou et al., 2019) and give way to feelings of eeriness (Rhim et al., 2022), in line with the “uncanny valley” phenomenon (Pizzi et al., 2023). That is anthropomorphism initially creates a sense of companionship, but it becomes uncomfortable when users begin to attribute a mind to the chatbot. This feeling of uncanniness diminishes the positive effects of anthropomorphism on perceived social presence (Rhim et al., 2022).

Another concern is that chatbots’ sophisticated interactive features, such as voice, gestures, and facial expressions (Tang et al., 2025), can elevate consumer expectations of the survey experience. Problems arise when there is a mismatch between these heightened expectations and the chatbot’s conversational capabilities. The more humanlike chatbots look, the more respondents expect them to perform like humans. When chatbots cannot maintain natural dialogue, offer flexible responses, or adapt appropriately to respondents’ needs, the perceived social presence might worsen rather than improve.

Instead of being inherently beneficial or problematic, the influence of humanization depends on task sensitivity (Hasler et al., 2013; Zhou et al., 2019) and the chatbot’s humanlikeness (Conrad et al., 2015; Hasler et al., 2013; Rhim et al., 2022; Tang et al., 2025; Xiao et al., 2020b) in the survey interaction. Future researchers should balance chatbot humanization, user comfort, and task sensitivity when designing surveys. For example, anthropomorphic design elements (Conrad et al., 2015; Hasler et al., 2013; Rhim et al., 2022; Tang et al., 2025; Xiao et al., 2020b), such as humanlike tone, voice, and gestures, can enhance perceived social presence and user engagement, but they can trigger eeriness when they exceed users’ expectations. A promising research direction would be to determine the optimal degree of humanization to maximize engagement without causing discomfort, while accounting for the moderating role of task sensitivity.

Theme 5: Effective Response Capability

Offering a critical capability in surveys, chatbots can generate higher response rates and superior response quality compared with traditional methods (Celino & Re Calegari, 2020; Guo et al., 2021; Hasler et al., 2013; Li et al., 2025; Soni et al., 2022; Ward & Pond, 2015; Zhou et al., 2019). In social media settings, both chatbots and online surveys help mitigate social desirability bias and thus respondents are more likely to share information when interacting with chatbots (Beam, 2023; Jacobsen et al., 2025; Kim et al., 2019; Tang et al., 2025; Xiao et al., 2019). Beam (2023) also notes that many respondents click on an online survey invitation, but very often do not start with the survey. In contrast, respondents appear more willing to spend time completing a survey if receiving an invitation from a chatbot (Celino & Re Calegari, 2020), due in part to their interactivity, which increases engagement (Abbas et al., 2021; Belhaj et al., 2021; Söderström et al., 2021; Xiao et al., 2020b).

The perceived anonymity of chatbots increases respondents’ willingness to disclose. However, the low sense of accountability may lead them to offer careless or deliberately fake answers (Li et al., 2025). Xiao et al. (2020a) show that once participants realize that a survey chatbot cannot adequately understand their responses, they may intentionally provide false information, thereby undermining response quality.

Current survey chatbots are far from flawless, and respondents may feel limited by their limited conversational abilities (Ramesh & Chawla, 2022). This may imply that participants are less willing to offer qualified answers. Contemporary NLP capabilities are not sophisticated enough to handle the nuances of human conversation (Hasler et al., 2013); they often struggle with slang, processing synonyms, and extracting entities. Although conversational surveys are designed to feel natural and engaging, limited NLP capabilities can disrupt conversational flow and frustrate respondents, ultimately reducing their willingness to provide thoughtful responses. Furthermore, a conversational approach may require more time and effort than participants initially anticipate, making it unclear whether they will complete all questions or provide in-depth responses (Ward & Pond, 2015; Zhou et al., 2019).

Future research is required to address current limitations in chatbot surveys. First, studies could examine respondents’ accountability to inspect the influence of their sense of responsibility on response accuracy (Li et al., 2025; Xiao et al., 2020a). Second, scholars could examine how respondents’ cognitive effort affects response time and willingness to complete the task (Zhou et al., 2019). Third, conversational design needs to balance natural conversations and limited natural language processing capabilities. Table 4 summarizes the main benefits and limitations, and corresponding mitigating factors across the five found themes.

Table 4.

Five Key Themes

Theme	Benefits	Limitations	Potentially mitigating factors
Data collection	• Enhance participant disclosure (Beam, 2023; Jin et al., 2024; Lucas et al., 2014; Te Pas et al., 2020; Zhou et al., 2019) • Improve response quality (Beam, 2023; Jin et al., 2024; Kim et al., 2019; Sidaoui et al., 2020) • Reduce satisficing behaviour (Hassina & Hiba, 2025)	• Generate frustration (Park et al., 2019) • Dissatisfaction (Jacobsen et al., 2025; Kim et al., 2019; Li et al., 2025) • Conversational disengagement (Hasler et al., 2013; Tang et al., 2025)	• Include structured conversations (Araujo, 2020; Jin et al., 2024; Ward & Pond, 2015) • Use minimal or repeated probing (Li et al., 2025; Park et al., 2019; Zarouali et al., 2024) • Use constrained responses (Jacobsen et al., 2025; Jin et al., 2024) • Use diverse open-ended responses • Avoid interaction breakdowns (Park et al., 2019; Xiao et al., 2019) • Respondents’ unfamiliarity with chatbots needs to be considered (Lucas et al., 2014; Tang et al., 2025; Zarouali et al., 2024)
Conversational interface	• Provides users with a natural, familiar way to express themselves (Celino & Re Calegari, 2020) • Efficiency (Belhaj et al., 2021; Pickard et al., 2016) • Highly flexible (Xiao et al., 2020a; 2020b) • Capture user attention and foster trust (Rhim et al., 2022; Tang et al., 2025)	• Unable to address specific user demands (Rhim et al., 2022) • Depend on pre-scripted conversational flows (Jacobsen et al., 2025)	• Participants behave conversationally (natural dialogue) or procedurally (following instructions) (Belhaj et al., 2021) • Aim to meet respondents’ expectations (Rhim et al., 2022) • Internet connectivity complicates the conversation
Interactivity	• Minimize survey fatigue (Hassina & Hiba, 2025) • Enhance response quality (Beam, 2023; Guo et al., 2021; Jacobsen et al., 2025; Kim et al., 2019) • Positive engagement (Belhaj et al., 2021; Celino & Re Calegari, 2020; Duka & Njeguš, 2021; Kim et al., 2019; Soni et al., 2022; Tang et al., 2025; Ward & Pond, 2015) • Favorable attitudes (Belhaj et al., 2021; Celino & Re Calegari, 2020; Duka & Njeguš, 2021)	• Exhibit disengagement (Gupta & Chen, 2022; Li et al., 2025) • Skip open-ended questions (Jacobsen et al., 2025; Zarouali et al., 2024) • Offer irrelevant or minimal answers (Xiao et al., 2020a; 2020b) • Spend longer completion time (Kim et al., 2019; Soni et al., 2022; Zarouali et al., 2024)	• Perceive chatbots as tools or partners in the interaction • Avoid a chatbot that is overly time-consuming (Rhim et al., 2022; Zarouali et al., 2024) • Avoid that the chatbot is cognitively demanding (Jacobsen et al., 2025)
Humanization	• Enhanced engagement (Conrad et al., 2015; Rhim et al., 2022; Xiao et al., 2020a, 2020b) • Willingness to respond (Conrad et al., 2015; Xiao et al., 2020b) • Motivate greater self-disclosure (Conrad et al., 2015; Hasler et al., 2013; Kim et al., 2019) • Provide interactive and personalized experiences (Kim et al., 2019; Rhim et al., 2022; Tang et al., 2025; Zarouali et al., 2024)	• Provoke discomfort or diminish positive feelings (Hasler et al., 2013; Kim et al., 2019; Zhou et al., 2019) • Cause feelings of eeriness (Rhim et al., 2022) • Decrease perceived social presence (Rhim et al., 2022)	• Task sensitivity (low vs. high) (Zhou et al., 2019) • Chatbot’s human likeness needs to be at an appropriate level (Conrad et al., 2015; Hasler et al., 2013; Rhim et al., 2022; Tang et al., 2025; Xiao et al., 2020b)
Effective response	• Share information (Beam, 2023; Jacobsen et al., 2025; Kim et al., 2019; Tang et al., 2025; Xiao et al., 2019) • Increase survey completion willingness (Beam, 2023; Celino & Re Calegari, 2020; Sidaoui et al., 2020) • Motivate engagement (Abbas et al., 2021; Belhaj et al., 2021; Söderström et al., 2021; Xiao et al., 2020b)	• Deliberately provide fake answers (Li et al., 2025) • Limited natural language processing ability (Hasler et al., 2013) • Need additional time and effort (Soni et al., 2022; Ward & Pond, 2015; Zhou et al., 2019)	• Respondents’ accountability to complete the survey (Li et al., 2025; Xiao et al., 2020a) • Minimize respondents’ cognitive effort (Zhou et al., 2019) • Natural or limited natural conversational design

Other Emerging Themes

We identified five emerging capabilities: chatbot conversational styles, gamification, response delay, chatbot questionnaire contextuality, and comfortable information sharing, as seen in Table 5. Although these themes have received limited attention in extant literature, with one or two supporting articles each, they offer significant potential to enhance the functionality, responsiveness, and recreational aspects of chatbots in online surveys. The themes may also contribute to defining future research avenues.

Table 5.

Five Emerging Themes

Five additional themes
Conversational styles	Casual styles (i.e., informal language, emojis, and abbreviations) lead to higher-quality responses to multiple-choice questions (Abbas et al., 2021).
Gamification	Gamification has a strong positive impact on survey completion (Te Pas et al., 2020).
Response delay	Effective methods for mitigating response delays by calculating response time based on response length, including the number of characters, sentence length, and average syllables per word (Rhim et al., 2022).
Chatbot questionnaire contextuality	Participants provide deeper and more reflective insights in a broader context of the questionnaire (Jacobsen et al., 2025).
Comfortable information sharing	The chatbot can adapt its questioning when it detects participants’ hesitation or distress, refraining from further probing unless participants indicate a willingness to continue (Tang et al., 2025).

Chatbot Conversational Styles

Different chatbot conversational styles might affect both user perceptions and response quality. For example, a study comparing reactions to a travel booking chatbot that uses modern versus Shakespearean language styles reveals that people find the former more user-friendly and the latter more entertaining (Elsholz et al., 2019). Abbas et al. (2021) demonstrate that a casual script, filled with emojis, results in higher-quality responses to multiple-choice questions than a standard survey script. A notable gap arises in this research stream, though, in that these studies typically use a terse, non-conversational baseline for comparison, rather than a formal yet conversational style.

Gamification

Incentives (financial or otherwise) can increase response and survey completion rates and have been used extensively. Survey chatbots can apply a more varied range of motivational elements to encourage survey completion. The introduction of nudging techniques, such as gamification, might be beneficial. As detailed in behavioral psychology, nudging involves gentle, non-forceful efforts to stimulate a desired action. Te Pas et al. (2020) report a strong positive impact of gamification on survey completion rates. If adding gamification makes chatbot surveys more pleasant to complete, it could also improve the quality of the survey results.

Response Delay

Response delays offer critical social cues in chatbot surveys. Unlike human discussants, who take time to read and formulate responses, retrieval-based chatbots can respond almost instantaneously. Immediate responses feel unnatural, thereby reducing user satisfaction relative to chatbots that use dynamic delays. Rhim et al. (2022) propose effective methods for adjusting response speed and mitigating response delays by calculating response time based on response length, including the number of characters, sentence length, and average syllables per word. A dynamically timed response can enhance perceptions of social presence and trustworthiness. Users are also likely to develop subjective interpretations of a delay.

Chatbot Questionnaire Contextuality

In a chatbot survey, the nuanced richness comes from the chat data. As with face-to-face interviews, the most valuable insights often do not emerge from individual participant statements alone, but from the broader context of the questionnaire, such as patterns in response timing, follow-up answers, or repeated clarifications. Jacobsen et al. (2025) observe that participants initially struggled to articulate their thoughts about complex topics, such as personal finance habits. However, as the chatbot guided them through step-by-step probing questions, such as asking about daily routines first, then exploring underlying motivations, participants gradually provided deeper and more reflective insights. This example underscores the potential of chatbots to uncover rich, context-dependent information in survey settings.

Comfortable Information Sharing

For participants, it often is challenging to discuss sensitive or personal topics, particularly when they feel judged or stressed. Retaining respect for participants’ autonomy in user-centered design underscores the significance of the boundaries they set, which can help create a comfortable or safe environment for disclosure. The survey chatbot TigerGPT (Tang et al., 2025) maintains a relaxed, supportive, and non-judgmental tone when interacting with participants, particularly when addressing sensitive questions. This chatbot can adapt its questioning when it detects participants’ hesitation or distress, refraining from further probing unless participants indicate a willingness to continue. By fostering a comfortable survey environment, this balanced method respects participants, encourages them to share information more openly, and helps preserve their sense of psychological safety.

Discussion

AI transforms the marketing research landscape into the collection and analysis of information, the execution of strategies (e.g., ad placement), or the evaluation of their impact. We seek to establish AI applications at an initial stage, namely, collecting consumer information through chatbot surveys. A thematic analysis of chatbot capabilities for online surveys identified five key themes, i.e., data collection, conversational interface, interactivity, humanization, and effective response, and five less commonly occurring themes, i.e., conversational styles, gamification, response delay, chatbot questionnaire contextuality, and comfortable information sharing, see Tables 4 and 5. We explored the benefits and limitations of chatbots in terms of each theme. The reported findings provide valuable insights for researchers and marketers seeking to improve survey design, data or feedback collection, respondent engagement, and response quality by integrating chatbots into their online surveys.

As a theoretical contribution, this review synthesizes prior work across disciplines, resulting in an organization of chatbot survey characteristics into five key themes and five emerging themes. These themes offer systematic lenses through which future studies can examine key chatbot survey outcomes. Attributes that can be evaluated are, for example, respondents’ disclosure willingness (Beam, 2023; Jacobsen et al., 2025; Jin et al., 2024), survey completion rates (Celino & Re Calegari, 2020), and respondent engagement (Belhaj et al., 2021).

The five key themes and the five emerging themes that we report and discuss extend theoretical understanding of human-AI interaction in survey contexts. Earlier research only focused on specific aspects relating to the use of chatbots for surveys. For instance, Höhne et al. (2024) focused on Gemini Pro; this review examines the broader use of general-purpose survey chatbots in data collection, while we synchronized the evidence about humanization of chatbots for conducting surveys. Furthermore, Dwivedi et al. (2021) addressed AI-related challenges and opportunities at a macro level; the literature review reported in our paper focuses on chatbot-enabled online surveys and empirically reinforces the role of “humanness” in shaping survey effectiveness, see also Rhim et al. (2022) and Xiao et al. (2020b). Our review highlights the need to establish the optimal level of chatbot humanization that enhances engagement without eliciting discomfort (Kim et al., 2019; Zhou et al., 2019) or perceived eeriness (Rhim et al., 2022). More generally, our review provides an overview of the broader range of themes addressed in previous papers. This oversight results in well-grounded managerial implications and opens novel avenues for research on how conversational design, interactional cues, and chatbot behavioral features shape respondents’ experiences.

We propose four main practical implications. First, the five reported key capabilities provide guidelines about online market research situations in which it is appropriate to deploy chatbots. Table 6 provides advice for marketers based on the five dimensions. For example, marketers could consider chatbots to be rather useful when the questionnaire objective needs a higher level of interactivity (Belhaj et al., 2021; Duka & Njeguš, 2021; Kim et al., 2019; Soni et al., 2022; Tang et al., 2025; Ward & Pond, 2015) or improved completion rates (Beam, 2023; Celino & Re Calegari, 2020), such as improved completion rate for sensitive topics (Kim et al., 2019; Zhou et al., 2019). We also found that chatbot surveys may be less effective for highly complex questionnaires (Zarouali et al., 2024).

Table 6.

Advice for Market Researchers and a Research Agenda

Theme	Advice for market researchers	Research agenda
Data collection	• Use constrained responses to minimize ambiguity and misinterpretation (Hasler et al., 2013; Jacobsen et al., 2025; Jin et al., 2024) • Adopt diverse open-ended responses regarding experience and preference (Xiao et al., 2020a; 2020b) • Make strategies to address respondents’ unfamiliarity with conversational survey formats to enhance response quality and reliability (Lucas et al., 2014; Tang et al., 2025; Zarouali et al., 2024)	• Explore the design of structured conversations in chatbot surveys to offer clear guidance and a sense of dialogue (Araujo, 2020; Jin et al., 2024; Ward & Pond, 2015) • Examine the effects of minimal or repeated probing (Li et al., 2025; Park et al., 2019; Zarouali et al., 2024) • Investigate how to predict and fix interaction breakdowns if chatbots misunderstand inputs (Park et al., 2019; Xiao et al., 2019)
Conversational interface	• Monitor how participants interact with chatbot surveys, distinguishing between conversational behaviors (e.g., natural dialogue) and procedural behaviors (e.g., following instructions) (Belhaj et al., 2021) • Combine chatbot capabilities with human supervision for potential internet or connectivity issues,	• Investigate respondents’ engagement with chatbots in the conversational interface • Examine whether the design of conversational interfaces meets respondents’ expectations (Rhim et al., 2022).
Interactivity	• Design the chatbot to fit respondents’ perception of interactivity • Balance interactivity and respondents’ cognitive overload (Soni et al., 2022; Zhou et al., 2019)	• Investigate how respondents perceive chatbots: Survey tools vs. conversational partners • Explore the effects of interactivity, including potential time-consuming (Rhim et al., 2022; Zarouali et al., 2024) or cognitively demanding (Jacobsen et al., 2025)
Humanization	• Use testing and iterative design to help identify the optimal level of humanization to maximize engagement and avoid the uncanny valley effect • Balance chatbot humanization, user comfort, and task sensitivity in survey design	• Examine anthropomorphic design elements (tone, voice, gestures) and their impact on engagement and social presence (Conrad et al., 2015; Hasler et al., 2013; Rhim et al., 2022; Tang et al., 2025; Xiao et al., 2020b) • Consider task sensitivity (Zhou et al., 2019) as a moderating factor influencing chatbot design effectiveness
Effective response	• Aware that the complexity of questions and interactions affects respondents’ survey experience (Beam, 2023; Celino & Re Calegari, 2020). Perhaps simplifying interactions without sacrificing information quality can improve engagement • Notice chatbots’ language processing limitations and design interaction limitations to minimize misunderstandings	• Investigate how participants’ sense of responsibility influences response accuracy (Li et al., 2025; Xiao et al., 2020a) • Examine how cognitive effort affects response time, completion rates, and willingness to participate (Zhou et al., 2019) • Explore improvements in chatbot interactions, balancing natural conversational flow with limitations in natural language processing capabilities

Second, we offer recommendations for deploying chatbots, including selecting and designing chatbot attributes based on the survey task and the target audience’s characteristics. Chatbots may capture user attention and may reduce survey attrition rates (Xiao et al., 2020a; 2020b). Nevertheless, when using chatbots for market research, it is key to limit unnecessary conversational features to minimize cognitive load (Soni et al., 2022; Zhou et al., 2019) and respondent fatigue (Hassina & Hiba, 2025). Managers should evaluate tone, interactivity, response length, and levels of personalization to ensure chatbots support the intended research objectives.

Third, we emphasize several design and implementation considerations. For example, adopting humanization cues (i.e., tone, persona, and social presence) can foster trust and rapport but requires careful calibration to avoid perceived discomfort or creepiness (Rhim et al., 2022). Creepiness occurs when the chatbot recalls and uses too many personal details, or when the conversation becomes so friendly that it is only suited to friends (Pizzi et al., 2023).

Fourth, we highlight several practical challenges that marketers need to anticipate and manage, summarized in Table 4 as mitigating factors. These include respondent fatigue caused by overly long or repetitive conversations (Park et al., 2019; Zarouali et al., 2024), a lack of human intervention or interaction breakdowns (Park et al., 2019), and misunderstandings of open-ended responses. Addressing such possible issues requires conducting several pretests, refining chatbot logic, and integrating fallback solutions, such as hybrid human-chatbot survey support.

This study has several limitations that suggest avenues for future research. Research into optimizing chatbot designs across diverse user demographics, technological settings, and survey purposes is critical. Studies might explore integrating advanced AI techniques, such as sentiment analysis (Eyu et al., 2025), NLP enhancements, and personalization, to enhance the effectiveness of chatbot surveys. As Hasal et al. (2021) recommend, we note the need for research to identify options for improving chatbot security and privacy protection across two key domains: secure data transfer and secure data storage. Ongoing research must account for the rapid evolution of generative AI and its potential to increase consumer acceptance of chatbots across contexts, including surveys. By addressing these methodological and ethical considerations, researchers could unlock the full potential of chatbots as reliable, scalable, and valuable tools for data collection and analysis in both academic and industry settings.

Supplemental Material

Supplemental Material - Chatbot-Administered Surveys in Marketing: A Literature Review

Supplemental Material for Chatbot-Administered Surveys in Marketing: A Literature Review by Yingxue Zhao, Xiaoyi (Sylvia) Gao, Leo Paas in International Journal of Market Research.

Footnotes

ORCID iDs

Yingxue Zhao

Xiaoyi (Sylvia) Gao

Leo Paas

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental Material

Supplemental material for this article is available online.

References

Abbas

Pickard

Atwell

Walker

(2021). University student surveys using chatbots: Artificial intelligence conversational agents. In: Learning and collaboration technologies: Games and virtual environments for learning. 8th International Conference, LCT 2021, held as part of the 23rd HCI International Conference, HCII 2021, Proceedings, Virtual Event, July 24–29, 2021, 12785, pp. 155–169.

Araujo

(2020). Conversational agent research toolkit. Computational Communication Research, 2(1), 35–51. https://doi.org/10.5117/CCR2020.1.002.ARAU

Beam

E. A.

(2023). Social media as a recruitment and data collection tool: Experimental evidence on the relative effectiveness of web surveys and chatbots. Journal of Development Economics, 162, 1–11. https://doi.org/10.1016/j.jdeveco.2023.103069

Belhaj

Hamdane

Chaoui

N. E. H.

Chaoui

El Bekkali

(2021). Engaging students to fill surveys using chatbots: University case study. Indonesian Journal of Electrical Engineering and Computer Science, 24(1), 473–483. https://doi.org/10.1016/j.jdeveco.2023.103069

Celino

Re Calegari

(2020). Submitting surveys via a conversational interface: An evaluation of user acceptance and approach effectiveness. International Journal of Human-Computer Studies, 139, 1–16. https://doi.org/10.1016/j.ijhcs.2020.102410

Conrad

F. G.

Schober

M. F.

Jans

Orlowski

R. A.

Nielsen

Levenstein

(2015). Comprehension and engagement in survey interviews with virtual agents. Frontiers in Psychology, 6, 1–20. https://doi.org/10.3389/fpsyg.2015.01578

Cruzes

D. S.

Dyba

(2011). Recommended steps for thematic synthesis in software engineering. In 2011 international symposium on empirical software engineering and measurement, Alberta, Canada, Sept 22-23 2011, pp. 275–284.

Downie

Hayes

(2025). A guide to AI customer service chatbots. IBM. https://www.ibm.com/think/topics/ai-customer-service-chatbots

Duka

Njeguš

(2021). Conversational survey chatbot: User experience and perception. In: Sinteza 2021-International scientific conference on information technology and data related research (pp. 322–327). https://portal.sinteza.singidunum.ac.rs/Media/files/2021/322-327.pdf

10.

Dwivedi

Y. K.

Hughes

Ismagilova

Aarts

Coombs

Crick

Duan

Dwivedi

Edwards

Eirug

Galanos

Ilavarasan

P. V.

Janssen

Jones

Kar

A. K.

Kizgin

Kronemann

Lal

Lucini

Williams

M. D.

(2021). Artificial intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice and policy. International Journal of Information Management, 57, 1–47. https://doi.org/10.1016/j.ijinfomgt.2019.08.002

11.

Dwyer

F. R.

Schurr

P. H.

(1987). Developing buyer-seller relationships. Journal of Marketing, 51(2), 11–27. https://doi.org/10.1177/002224298705100202

12.

Elsholz

Chamberlain

Kruschwitz

(2019). Exploring language style in chatbots to increase perceived product value and user engagement. In Proceedings of the 2019 conference on human information interaction and retrieval, Glasgow, Scotland, UK, March 10-14, 2019, pp. 301–305.

13.

Eyu

J. M.

Yau

K.-L. A.

Liu

Chong

Y.-W.

(2025). Reinforcement learning in sentiment analysis: A review and future directions. Artificial Intelligence Review, 58(1), 1–40. https://doi.org/10.1007/s10462-024-10967-0

14.

Gaikwad

Khadse

Wailthare

Dubey

(2018). Artificial intelligence-based chatbot. International Journal for Research in Applied Science and Engineering Technology, 6(4), 2305–2306. https://doi.org/10.22214/ijraset.2018.4393

15.

Guo

Zong

Chen

Zhao

Shang

(2021). Design and implementation of a new serverless conversational survey system. In 2021 IEEE international conference on data science and computer application (ICDSCA) (pp. 358–363), IEEE.

16.

Gupta

Chen

(2022). Supporting inclusive learning using chatbots? A chatbot-led interview study. Journal of Information Systems Education, 33(1), 98–108. https://jise.org/Volume33/n1/JISE2022v33n1pp98-108.html

17.

Hasal

Nowaková

Ahmed Saghair

Abdulla

Snášel

Ogiela

(2021). Chatbots: Security, privacy, data protection, and social aspects. Concurrency and Computation: Practice and Experience, 33(19), 1–13. https://doi.org/10.1002/cpe.6426

18.

Hasler

B. S.

Tuchman

Friedman

(2013). Virtual research assistants: Replacing human interviewers by automated avatars in virtual worlds. Computers in Human Behavior, 29(4), 1608–1616. https://doi.org/10.1016/j.chb.2013.01.004

19.

Hassina

Hiba

Malika

(2025). Designing intelligent electronic questionnaires through artificial intelligence: Enhancing respondent motivation and user experience in academic field research. Science, Education and Innovations in the Context of Modern Problems, 10(8), 1–12. https://doi.org/10.56334/sei/8.10.94

20.

Höhne

J. K.

Claassen

Shahania

Broneske

(2024). Bots in web survey interviews: A showcase. International Journal of Market Research, 0(0), 1–10. https://doi.org/10.1177/14707853241297009

21.

Jacobsen

R. M.

Cox

S. R.

Griggio

C. F.

van Berkel

(2025). Chatbots for data collection in surveys: A comparison of four theory-based interview probes. In: Proceedings of the 2025 CHI conference on human factors in computing systems, Yokohama Japan, 26 April-1 May 2025, pp. 1–21.

22.

Jin

Chen

Zhao

Cai

(2024). The way you assess matters: User interaction design of survey chatbots for mental health. International Journal of Human-Computer Studies, 189, 1–17. https://doi.org/10.1016/j.ijhcs.2024.103290

23.

Kim

Lee

Gweon

(2019). Comparing data from chatbot and web surveys: Effects of platform and conversational style on survey response quality. In: Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, Glasgow Scotland Uk, May 4-9 2019, pp. 1–12.

24.

Tauseef

Truong

K. N.

Mariakakis

(2025). A comparative analysis of information gathering by chatbots, questionnaires, and humans in clinical pre-consultation. In: Proceedings of the 2025 CHI conference on human factors in computing systems, Yokohama Japan, 26 April-1 May 2025, pp. 1–17.

25.

Lucas

G. M.

Gratch

King

Morency

L. P.

(2014). It’s only a computer: Virtual humans increase willingness to disclose. Computers in Human Behavior, 37, 94–100. https://doi.org/10.1016/j.chb.2014.04.043

26.

Meline

(2006). Selecting studies for systemic review: Inclusion and exclusion criteria. Contemporary Issues in Communication Science and Disorders, 33(Spring), 21–27. https://doi.org/10.1044/cicsd_33_s_21

27.

Neuendorf

K. A.

(2018). Content analysis and thematic analysis. In: Advanced research methods for applied psychology: Design, analysis and reporting (pp. 211–223). Taylor & Francis.

28.

Page

M. J.

Page

J. E.

McKenzie

P. M.

Bossuyt

Boutron

T. C.

Hoffmann

C. D.

Mulrow

McGuinness

L. A.

Akl

E. A.

Brennan

S. E.

Chou

Glanville

Grimshaw

J. M.

Hróbjartsson

Lalu

M. M.

Loder

E. W.

Mayo-Wilson

McDonald

Moher

(2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. British Medical Journal, 372(71), 1–9. https://doi.org/10.1136/bmj.n71

29.

Park

Choi

Lee

Kim

Lee

(2019). Designing a chatbot for a brief motivational interview on stress management: Qualitative case study. Journal of Medical Internet Research, 21(4), 1–13. https://doi.org/10.2196/12231

30.

Patino

C. M.

Ferreira

J. C.

(2018). Inclusion and exclusion criteria in research studies: definitions and why they matter. Jornal Brasileiro de Pneumologia, 44(2), 84–84. https://doi.org/10.1590/S1806-37562018000000088

31.

Pickard

M. D.

Roster

C. A.

Chen

(2016). Revealing sensitive information in personal interviews: Is self-disclosure easier with humans or avatars and under what conditions? Computers in Human Behavior, 65, 23–30. https://doi.org/10.1016/j.chb.2016.08.004

32.

Pizzi

Vannucci

Mazzoli

Donvito

(2023). I, chatbot! the impact of anthropomorphism and gaze direction on willingness to disclose personal information and behavioral intentions. Psychology and Marketing, 40(7), 1372–1387. https://doi.org/10.1002/mar.21813

33.

Ramesh

Chawla

(2022). Chatbots in marketing: A literature review using morphological and co-occurrence analyses. Journal of Interactive Marketing, 57(3), 472–496. https://doi.org/10.1177/10949968221095549

34.

Rhim

Kwak

Gong

Gweon

(2022). Application of humanization to survey chatbots: Change in chatbot perception, interaction experience, and survey data quality. Computers in Human Behavior, 126, 1–12. https://doi.org/10.1016/j.chb.2021.107034

35.

Sidaoui

Jaakkola

Burton

(2020). AI feels you: Customer experience assessment via chatbot interviews. Journal of Service Management, 31(4), 745–766. https://doi.org/10.1108/JOSM-11-2019-0341

36.

Smutny

Schreiberova

(2020). Chatbots for learning: A review of educational chatbots for the Facebook messenger. Computers and Education, 151, 1–11. https://doi.org/10.1016/j.compedu.2020.103862

37.

Söderström

Shatte

Fuller-Tyszkiewicz

(2021). Can intelligent agents improve data quality in online questionnaires? A pilot study. Behavior Research Methods, 53(5), 2238–2251. https://doi.org/10.3758/s13428-021-01574-w

38.

Soni

Ivanova

Wilczewski

Bailey

Ong

Narma

Bunnell

B. E.

Welch

B. M.

(2022). Virtual conversational agents versus online forms: Patient experience and preferences for health data collection. Frontiers in Digital Health, 4, 1–13. https://doi.org/10.3389/fdgth.2022.954069

39.

Tang

Chen

Shang

(2025). TigerGPT: A new AI chatbot for adaptive campus climate surveys. arXiv preprint arXiv:2504.13925. https://arxiv.org/pdf/2504.13925

40.

Te Pas

M. E.

Rutten

W. G. M. M.

Bouwman

R. A.

Buise

M. P.

(2020). User experience of a chatbot questionnaire versus a regular computer questionnaire: Prospective comparative study. JMIR Medical Informatics, 8(12), 1–10. https://doi.org/10.2196/21982

41.

Tsai

W. H. S.

Liu

Chuan

C. H.

(2021). How chatbots' social presence communication enhances consumer engagement: The mediating role of parasocial interaction and dialogue. The Journal of Research in Indian Medicine, 15(3), 460–482. https://doi.org/10.1108/JRIM-12-2019-0200

42.

Tyumeneva

(2026). Chatbot statistics 2025–2026: Market size, usage & trends. Elfsight. https://elfsight.com/blog/chatbot-statistics-and-trends/

43.

Ward

Pond

S. B.

(2015). Using virtual presence and survey instructions to minimize careless responding on internet-based surveys. Computers in Human Behavior, 48, 554–568. https://doi.org/10.1016/j.chb.2015.01.070

44.

Wohlin

Mendes

Felizardo

K. R.

Kalinowski

(2020). Guidelines for the search strategy to update systematic literature reviews in software engineering. Information and Software Technology, 127, 1–12. https://doi.org/10.1016/j.infsof.2020.106366

45.

Xiao

Zhou

M. X.

Chen

Yang

Chi

(2020a). If I hear you correctly: Building and evaluating interview chatbots with active listening skills. In: Proceedings of the 2020 CHI conference on human factors in computing systems, Honolulu, HI, USA, April 25-30 2020, pp. 1–14.

46.

Xiao

Zhou

M. X.

W. T.

(2019). Who should be my teammates: Using a conversational agent to understand individuals and help teaming. In: Proceedings of the 24th international conference on intelligent user interfaces, Marina del Ray, CA, USA, March 17-20, 2019, pp. 437–447.

47.

Xiao

Zhou

M. X.

Liao

Q. V.

Mark

Chi

Chen

Yang

(2020b). Tell me about yourself: Using an AI-powered chatbot to conduct conversational surveys with open-ended questions. ACM Transactions on Computer-Human Interaction, 27(3), 1–37. https://doi.org/10.1145/3381804

48.

Zarouali

Araujo

Ohme

de Vreese

(2024). Comparing chatbots and online surveys for (longitudinal) data collection: An investigation of response characteristics, data quality, and user evaluation. Communication Methods and Measures, 18(1), 72–91. https://doi.org/10.1080/19312458.2022.2156489

49.

Zhou

M. X.

Mark

Yang

(2019). Trusting virtual agents: The effect of personality. ACM Transactions on Interactive Intelligent Systems, 9(2-3), 1–36. https://doi.org/10.1145/323207

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB