Authors: Aravindh Pranav.v, Dr. P. Vinodhini
Abstract: Artificial intelligence is changing the way businesses communicate with customers, particularly through AI voice and text agents used for customer support. These agents can respond quickly, operate continuously, handle large numbers of conversations, and support businesses across channels such as WhatsApp, websites, in-app chat, and voice calls. At the same time, customer acceptance depends on familiarity, comprehension, trust, tone, and the availability of human support. This study examines the adoption of AI voice and text agents and their perceived impact on business operations from the customer perspective, using the Karta AI internship context as its grounding. A descriptive, quantitative and cross-sectional research design was used. Primary data were collected through a structured questionnaire from 30 respondents who had interacted with AI voice or text-based customer support agents, using convenience sampling. Secondary information was obtained from academic literature, industry reports and company publications. The collected data were analysed using frequency and percentage analysis, Chi-Square tests and One-Way ANOVA. The findings show that 70% of respondents were aged 18–25, 60% stated that AI understood their queries correctly, and 56.67% supported increased use of AI agents in customer support. WhatsApp bots were the most commonly used channel at 36.67%. Financial/payment disputes, complex technical issues and emotionally sensitive complaints were the situations in which customers most preferred human support. Robotic or impersonal responses and difficulty reaching a human agent were the most frequently reported challenges. Statistical analysis found a significant relationship between familiarity with AI and willingness to see greater AI use in the future (p = 0.022), while no significant relationship was found between frequency of AI use and query comprehension (p = 0.823) or between age group and future outlook (p = 0.220). The study concludes that AI adoption is broadly welcomed, but long-term effectiveness depends on better comprehension, more human-like interaction, transparent escalation and a balanced human-AI support model.
DOI: http://doi.org/10.5281/zenodo.21887911
