Low response rates have always been a challenge in online surveys. Filling out a survey, especially those with open-ended questions, can become tedious for respondents. To address these issues, we propose a new approach that reduces the burden of typing responses by integrating traditional search algorithms and machine learning algorithms. A hybrid Trie-based search algorithm was introduced for word auto-completion, significantly reducing the number of keystrokes required for responses. Domain-specific survey data was input into the training model, demonstrating promising results in initial testing.