π€ Empower your CX team with open-source AI-powered response auto-drafting. π€
An open source library that lets you leverage LLMs and latest advancements in AI to automate customer support interactions. By connecting a Large Language Model (LLM) to your knowledge base and historical support tickets via embeddings & vector searching, you can accurately auto-draft responses to all customer requests.
You can use cx-copilot to auto-draft responses in Helpdesk, Intercom, Support Inboxes, Zendesk, and anywhere else you store & respond to customer requests.
The basis of cx-copilot is embedding, vector storing and vector searching. Vector embeddings are a way to represent text as a series of numbers in such a way that you can perform mathematical operations, such as similarity comparison. By first embedding all previous historical customer request tickets using an embedding model (like text-embedding-ada-002 from OpenAI) and storing the embeddings & the paired response from your company in a vector database, you can then perform a vector search for incoming support tickets, returning the closest-matching tickets based on cosine similarity. The final step is to prompt a Large Language Model (LLM) with your team's responses to the closest-matching historical tickets, generating an auto-drafted response which will answer your customerβs query while conforming to your tone & formatting tendencies.
Integrations | |
---|---|
Helpscout | |
Gmail | |
Intercom (coming soon) | |
Zendesk (coming soon) | |
Discord (coming soon) |
Join the Discord community for cx-copilot for support & project updates.