We Replaced Our Support Chatbot With One That Reads Docs

In the world of customer support, the frustration of interacting with a chatbot that confidently provides incorrect answers is all too familiar. These bots often rely on pre-programmed responses or limited data sets, leading to a poor user experience. However, SoftSolz has introduced a game-changing solution: a support chatbot that actually reads and understands your documentation.

What Makes SoftSolz's Chatbot Different?

Unlike traditional chatbots that are merely bolted onto a help center, SoftSolz's chatbot is integrated with a retrieval layer that allows it to access and understand your entire knowledge base. This is achieved through a sophisticated process where documents are parsed, chunked, embedded, and stored in pgvector. The chatbot's answers are grounded against selected knowledge base collections, ensuring accuracy and relevance.

How Does SoftSolz Ensure Data Security?

SoftSolz takes data security seriously. The embed widget is knowledge base-scoped, meaning a public widget can never leak beyond the collections you have bound to that key. This ensures that sensitive information remains protected while the chatbot provides accurate support.

Comparing SoftSolz with Intercom Fin and Zendesk AI

While Intercom Fin and Zendesk AI offer strong answer engines, they come with limitations that can be challenging for small teams. These platforms price per resolution and are confined within their own support suites. As a result, they cannot answer questions about your invoices or form submissions.

In contrast, SoftSolz's chatbot sits on top of the same workspace as your other services. This means the same assistant can access and understand both your knowledge base and product data, providing a more comprehensive support experience.

The Importance of Document Quality

It's important to note that the quality of the chatbot's retrieval is directly linked to the quality of the documents it accesses. SoftSolz provides a streamlined ingestion flow to ensure that your documents are properly formatted and organized for optimal chatbot performance.

Understanding the Ingestion Flow

The ingestion flow involves several key steps:

1. Document Parsing: Breaking down documents into manageable chunks.
2. Embedding: Converting text into a format that the chatbot can understand.
3. Storage: Saving the embedded data in pgvector for quick retrieval.
4. Grounding: Ensuring answers are based on the most relevant knowledge base collections.

By following this process, SoftSolz ensures that its chatbot provides accurate and reliable support.

Conclusion: Join the Waitlist

SoftSolz's innovative approach to support chatbots offers a significant improvement over traditional solutions. By grounding answers in your knowledge base and integrating with your existing services, it provides a more accurate and comprehensive support experience. To learn more, visit the Knowledge Hub product guide and join the waitlist today.