Table 1 Summary of papers Authors (doi)... | Emerald Publishing

Table 1

Summary of papers

Authors (doi)	Paper type (Qual, RCT, Quasi-Experimental design)	Sample characteristics	n (number of participants)	AI coaching delivery	AI coaching training	Results/Findings/Insight
Ellis-Brush (2021), https://doi.org/10.24384/er2p-4857	Mixed-methods: Quasi-experimental design and qualitative	Banking sector No age reported	48	Text-based	Cognitive behavioural therapy	Improved self-resilience, No significant Working alliance
Figueroa et al. (2021), https://doi.org/10.3389/fdgth.2021.747153	Qualitative	Low-income women, aged 27–41, majority Hispanic/Latine	18	Text-based	Behavioural activation, motivational interviewing, acceptance and commitment, and solution-focused therapy	Positive perception of the chatbot, showed interest in using chatbots for health improvements, concerns about data privacy
Hassoon et al. (2021), https://doi.org/10.1038/s41746-021-00539-9	Randomised control Trial (RCT)	Overweight or obese cancer survivors, mean age 62.1 years, 90% female, various cancer types	42	Text-based Voice-based	Physical activity interventions	Improved step count by voice bot compared to text bot and control, no significant difference between text bot and control
Kannampallil et al. (2022), https://doi.org/10.2196/38092	Mixed methods (Observational): Quantitative and qualitative	Patients with mild to moderate depression and/or anxiety, mean age 43.9 years, 77% female, 73% racial or ethnic minorities	26	Voice-based	Problem-Solving treatment	High pragmatic usability and favourable user experience, higher temporal workload during a problem-solving session
Mai et al. (2021), https://doi.org/10.1007/978-3-030-90328-2_29	Mixed-methods: Quasi-experimental design and qualitative	Students, aged 21–39, majority male (75%)	12	Text-based	Exam anxiety, solution-focused coaching	Disclosure to the chatbot and rapport, more self-disclosure and rapport found in the chatbot informational disclosure versus self-disclosure
Mai et al. (2022), https://doi.org/10.1038/s41746-021-00539-9	Exploratory quantitative study	University students	21	Conversational AI (writing), Rule-based (clicking)	Exam anxiety, solution-focused coaching	Moderate to high working alliance, higher value observed for bonding in the conversational AI, coaching through chatbots well accepted
Movsumova et al. (2020), https://doi.org/10.15219/em86.1485	Mixed-methods (qualitative and quantitative)	Varied demographics, including men and women of different ages, occupations, and positions	33	AI-based tool (Mentorbot) through Telegram, assisting human coaches in dialogue and session analysis	No specific training details for coaches on Mentorbot mentioned	Positive dynamics clarity, willingness to act and stress reduction. Mentorbot was effective for novel and confidential requests, while human coaches were stronger in reducing stress and perceived overall usefulness
Passmore et al. (2021), https://doi.org/10.53841/bpstcp.2021.17.2.41	Quantitative: Survey	Coaches from 79 countries, average age 54, 66% female	1200	N/A	N/A	Mixed views of the role of AI in coaching, equally divided seen as providing benefits and disbenefits
Passmore and Tee (2023), https://doi.org/10.1108/JWgAM-06-2023-0057	Cross-sectional, mixed-method study	Experts in coaching, academic program directors, experience in reviewing and marking coaching assignments	14	Text-based	Various prompts to evaluate GPT-4’s ability to define coaching, compare ethical codes, summarise meta-analyses, and conduct coaching	GPT-4 is capable of generating plausible content but often contains inaccuracies and falsified information Concerns over CPT-4 ethical judgement were highlighted
Stephens et al. (2019), https://doi.org/10.1093/tbm/ibz043	Feasibility study	Youth enrolled in a weight management program, mean age 15.2 years, 57% female, 43% Hispanic	23	Text-based	Behavioural coaching	AI coach was feasible and helpful; high engagement (4,123 messages), 96% found it useful, 81% reported positive progress toward goals
Terblanche and Cilliers (2020), https://doi.org/10.22316/poc/05.1.06	Exploratory study: Survey	Online participants, no age or demographics reported	226	Text-based	Goal-attainment theory, GROW model	Performance expectancy, social influence, and attitude significantly influence behavioural intent to use AI coach
Terblanche et al. (2022a), https://doi.org/10.24384/5cgf-ab69	Longitudinal RCT	Undergraduate students, diverse demographics, average age 22 years	168	Text-based	Goal-attainment theory, GROW model	Improved goal attainment, no significant changes in psychological well-being, resilience, or perceived stress
Terblanche et al. (2022b), https://doi.org/10.1371/journal.pone.0270255	Longitudinal RCT	Business school students, diverse demographics	478	Text-based	Goal-attainment theory, GROW model	Improved goal attainment compared to the control group, same effect on goal attainment as human coaches
Terblanche et al. (2023), https://doi.org/10.1080/17521882.2022.2094278	Qualitative study	Final year undergraduate students, aged 20–22, diverse cultures, low socioeconomic background	31	Text-based	Goal-setting	Positive attitude and performance expectations promoted engagement; AI coach perceived as accessible, easy to use, and intelligent; minimal perceived risk; social influence and information about the AI coach influenced adoption
Terblanche et al. (2024), https://doi.org/10.1080/17521882.2024.2304792	Qualitative study	Coaches and clients from a financial services organisation	16 (9 coaches, 7 clients)	Text-based	GROW model	Coaches were concerned about potential negative interference with the coach-client bond, while clients found the chatbot useful for goal tracking, accountability, and convenience. Clients felt psychologically safe with the chatbot and appreciated its non-judgmental nature
Tropeg et al. (2019), https://doi.org/10.2196/12805	Scoping review	Mostly within medical care	Not applicable (Review of 49 studies)	Text, voice and avatars, ECA	Various methods, focusing on health improvement	Effectively support physical activity and weight management. ECA have significant potential in promoting healthy behaviours

Authors (doi)	Paper type (Qual, RCT, Quasi-Experimental design)	Sample characteristics	n (number of participants)	AI coaching delivery	AI coaching training	Results/Findings/Insight
Ellis-Brush (2021),https://doi.org/10.24384/er2p-4857	Mixed-methods: Quasi-experimental design and qualitative	Banking sectorNo age reported	48	Text-based	Cognitive behavioural therapy	Improved self-resilience, No significantWorking alliance
Figueroa et al. (2021), https://doi.org/10.3389/fdgth.2021.747153	Qualitative	Low-income women, aged 27–41, majority Hispanic/Latine	18	Text-based	Behavioural activation, motivational interviewing, acceptance and commitment, and solution-focused therapy	Positive perception of the chatbot, showed interest in using chatbots for health improvements, concerns about data privacy
Hassoon et al. (2021), https://doi.org/10.1038/s41746-021-00539-9	Randomised control Trial (RCT)	Overweight or obese cancer survivors, mean age 62.1 years, 90% female, various cancer types	42	Text-basedVoice-based	Physical activity interventions	Improved step count by voice bot compared to text bot and control, no significant difference between text bot and control
Kannampallil et al. (2022), https://doi.org/10.2196/38092	Mixed methods (Observational): Quantitative and qualitative	Patients with mild to moderate depression and/or anxiety, mean age 43.9 years, 77% female, 73% racial or ethnic minorities	26	Voice-based	Problem-Solving treatment	High pragmatic usability and favourable user experience, higher temporal workload during a problem-solving session
Mai et al. (2021), https://doi.org/10.1007/978-3-030-90328-2_29	Mixed-methods: Quasi-experimental design and qualitative	Students, aged 21–39, majority male (75%)	12	Text-based	Exam anxiety, solution-focused coaching	Disclosure to the chatbot and rapport, more self-disclosure and rapport found in the chatbot informational disclosure versus self-disclosure
Mai et al. (2022), https://doi.org/10.1038/s41746-021-00539-9	Exploratory quantitative study	University students	21	Conversational AI (writing), Rule-based (clicking)	Exam anxiety, solution-focused coaching	Moderate to high working alliance, higher value observed for bonding in the conversational AI, coaching through chatbots well accepted
Movsumova et al. (2020), https://doi.org/10.15219/em86.1485	Mixed-methods (qualitative and quantitative)	Varied demographics, including men and women of different ages, occupations, and positions	33	AI-based tool (Mentorbot) through Telegram, assisting human coaches in dialogue and session analysis	No specific training details for coaches on Mentorbot mentioned	Positive dynamics clarity, willingness to act and stress reduction. Mentorbot was effective for novel and confidential requests, while human coaches were stronger in reducing stress and perceived overall usefulness
Passmore et al. (2021), https://doi.org/10.53841/bpstcp.2021.17.2.41	Quantitative: Survey	Coaches from 79 countries, average age 54, 66% female	1200	N/A	N/A	Mixed views of the role of AI in coaching, equally divided seen as providing benefits and disbenefits
Passmore and Tee (2023),https://doi.org/10.1108/JWgAM-06-2023-0057	Cross-sectional, mixed-method study	Experts in coaching, academic program directors, experience in reviewing and marking coaching assignments	14	Text-based	Various prompts to evaluate GPT-4’s ability to define coaching, compare ethical codes, summarise meta-analyses, and conduct coaching	GPT-4 is capable of generating plausible content but often contains inaccuracies and falsified informationConcerns over CPT-4 ethical judgement were highlighted
Stephens et al. (2019), https://doi.org/10.1093/tbm/ibz043	Feasibility study	Youth enrolled in a weight management program, mean age 15.2 years, 57% female, 43% Hispanic	23	Text-based	Behavioural coaching	AI coach was feasible and helpful; high engagement (4,123 messages), 96% found it useful, 81% reported positive progress toward goals
Terblanche and Cilliers (2020), https://doi.org/10.22316/poc/05.1.06	Exploratory study: Survey	Online participants, no age or demographics reported	226	Text-based	Goal-attainment theory, GROW model	Performance expectancy, social influence, and attitude significantly influence behavioural intent to use AI coach
Terblanche et al. (2022a), https://doi.org/10.24384/5cgf-ab69	Longitudinal RCT	Undergraduate students, diverse demographics, average age 22 years	168	Text-based	Goal-attainment theory, GROW model	Improved goal attainment, no significant changes in psychological well-being, resilience, or perceived stress
Terblanche et al. (2022b), https://doi.org/10.1371/journal.pone.0270255	Longitudinal RCT	Business school students, diverse demographics	478	Text-based	Goal-attainment theory, GROW model	Improved goal attainment compared to the control group, same effect on goal attainment as human coaches
Terblanche et al. (2023), https://doi.org/10.1080/17521882.2022.2094278	Qualitative study	Final year undergraduate students, aged 20–22, diverse cultures, low socioeconomic background	31	Text-based	Goal-setting	Positive attitude and performance expectations promoted engagement; AI coach perceived as accessible, easy to use, and intelligent; minimal perceived risk; social influence and information about the AI coach influenced adoption
Terblanche et al. (2024),https://doi.org/10.1080/17521882.2024.2304792	Qualitative study	Coaches and clients from a financial services organisation	16 (9 coaches, 7 clients)	Text-based	GROW model	Coaches were concerned about potential negative interference with the coach-client bond, while clients found the chatbot useful for goal tracking, accountability, and convenience. Clients felt psychologically safe with the chatbot and appreciated its non-judgmental nature
Tropeg et al. (2019),https://doi.org/10.2196/12805	Scoping review	Mostly within medical care	Not applicable (Review of 49 studies)	Text, voice and avatars, ECA	Various methods, focusing on health improvement	Effectively support physical activity and weight management. ECA have significant potential in promoting healthy behaviours

Source(s): Authors’ own