Home / Companies / Elastic / Blog / Post Details
Content Deep Dive

GenAI for customer support — Part 1: Building our proof of concept

Blog post from Elastic

Post Details
Company
Date Published
Author
-
Word Count
2,184
Language
-
Hacker News Points
-
Summary

Elastic's blog post details their journey in developing a generative AI application to enhance customer support, showcasing the creation of a proof of concept utilizing Vertex AI integrated with Salesforce Service Cloud. The initiative aims to automate case summaries and draft initial responses to improve efficiency and customer satisfaction while easing the workload on support agents. Initial user feedback highlighted a need for improved AI accuracy, particularly in handling technical queries due to limitations in the language model's training data. Consequently, Elastic plans to refine input data and set higher accuracy thresholds to enhance AI performance, with future iterations aiming to integrate Elasticsearch for better response precision. The ultimate goal is to develop a scalable Support AI Chat Assistant, facilitating a self-service experience that could reduce response times and enhance customer satisfaction, while also freeing up resources for more strategic tasks. The ongoing blog series promises further updates on this project, emphasizing Elastic's focus on customer-centric solutions.