Prior work on sentiment analysis using weak supervision primarily focuses on\ndifferent reviews such as movies (IMDB), restaurants (Yelp), products\n(Amazon).~One under-explored field in this regard is customer chat data for a\ncustomer-agent chat in customer support due to the lack of availability of free\npublic data. Here, we perform sentiment analysis on customer chat using weak\nsupervision on our in-house dataset. We fine-tune the pre-trained language\nmodel (LM) RoBERTa as a sentiment classifier using weak supervision. Our\ncontribution is as follows:1) We show that by using weak sentiment classifiers\nalong with domain-specific lexicon-based rules as Labeling Functions (LF), we\ncan train a fairly accurate customer chat sentiment classifier using weak\nsupervision. 2) We compare the performance of our custom-trained model with\noff-the-shelf google cloud NLP API for sentiment analysis. We show that by\ninjecting domain-specific knowledge using LFs, even with weak supervision, we\ncan train a model to handle some domain-specific use cases better than\noff-the-shelf google cloud NLP API. 3) We also present an analysis of how\ncustomer sentiment in a chat relates to problem resolution.\n