AAI_2025_Capstone_Chronicles_Combined

‭ResolveAI‬

‭uniform input for the model. Overall, these steps ensure that the textual data is cleaned,‬ ‭normalized, and prepared effectively for the downstream machine learning tasks.‬

‭Exploratory Data Analysis‬

‭This combined histogram (Fig. 1) contrasts the character lengths of customer message‬ ‭bodies (blue) versus agent answers (orange). Overall, answer lengths tend to be longer,‬

‭particularly in the 200–500 character range, suggesting that support agents often provide more‬ ‭detailed information than what customers include in their initial inquiries. While both‬ ‭distributions exhibit peaks in the lower hundreds of characters (indicating relatively concise‬ ‭conversations) the agent responses maintain a noticeable tail at higher character counts. This‬ ‭implies that certain inquiries require more comprehensive explanations or troubleshooting steps,‬ ‭prompting agents to write extended responses.‬

‭7‬

55