Making AI useful in our everyday lives is a company mission at ClearFeed. Today - we’re thrilled to introduce significant additions to our GPT-Powered Answers functionality:
- Introducing ClearFeed Virtual Agent.
- Connectors for Notion, PDF files, Zendesk & FreshDesk KBs.
- Significant enhancements to search results & answer quality.
With ClearFeed's Virtual Agent - you can now deflect common customer queries with GPT-powered answers without any agent intervention. The enhanced capability builds on our earlier release of AI-assisted agent support. Support teams can now automate support interactions, improving response time, and providing both agents and customers with useful information in the context of a customer query.
Let’s dive deep into how you can leverage ClearFeed’s Virtual Agent and other enhancements for next-gen Slack-Based Customer and Employee Support.
How the Virtual Agent works
The Virtual Agent is designed to operate within Slack request channels, acting as the first point of contact for customer queries. It uses advanced algorithms to sift through your indexed support content, including knowledge bases, documentation, and FAQs, to provide immediate, relevant responses.
Instant Response with GPT-Powered Answers
Upon receiving a customer query on Slack, the Virtual Agent provides a GPT-generated response, aiming for resolution without agent intervention. This system is designed to deflect tickets by answering questions directly through your documented knowledge.
Each generated answer includes References - documents directly used to construct the answer, providing users with a transparent look at the source material. Additionally, Search Results offer links to additional, relevant information.
Integration with Private and Public Knowledge Sources
The Virtual Agent leverages your existing documentation, FAQs, and wikis, ensuring that answers are instant, relevant, and accurate. ClearFeed supports various knowledge sources including Zendesk Guides, Notion, Freshdesk, Confluence, and more.
This deep integration allows for a seamless user experience, as responses are directly aligned with your organization's knowledge base.
Additional read: Learn how to index your Knowledge Sources, and activate the Virtual Agent in your Slack channels.
Configurable Answering Modes
You can tailor the Virtual Agent to either provide full GPT-generated answers with documentation references or to limit responses to search results, accommodating varying levels of AI trust among organizations. You can configure either of the two answering modes:
- Search Results Only: The Virtual Agent simply posts these snippets (as Search Results) in a threaded reply to the request.
- Generated Answers and Search Results: The Virtual Agent uses GPT to generate an answer using the most relevant snippet. Then, it posts this Generated Answer, along with the snippets (as Search Results) in a threaded reply to the request.
Automatic Ticket Creation for Unresolved Queries
When queries cannot be resolved through automated responses, the Virtual Agent works together with your ticketing platform (ClearFeed ticketing or external ticketing) on Slack to seamlessly direct these cases to human support by creating support tickets. This is how the system works under different scenarios:
- When an answer is generated successfully: Users are presented with a "No, create a ticket" option if the answer doesn't resolve their query, alongside a "Yes, mark as solved" for satisfactory responses. Clicking "No" collapses the answer and initiates ticket creation.
- When no answer is generated: If the Virtual Agent cannot generate an answer, a ticket is automatically created to ensure the query is timely addressed by a support agent.
- When no Feedback is provided by the customer: If no feedback is given within 1 hour after an answer is provided, ticket creation is automatically triggered to capture unresolved queries. This time window can be customized based on your requirements.
These mechanisms ensure efficient resolution of complex issues while maintaining a high level of support quality.
Improvements to Search and Answer Quality
We are also excited to announce several improvements to our AI infrastructure to provide users with better answers and search results:
- Use of Cohere Embeddings: We have recently migrated from OpenAI’s ada-002 embeddings to Cohere to improve retrieval quality.
- Answer Verification Module: Answers generated by OpenAI’s GPT-4 are now re-verified by another AI agent to reduce hallucinations.
- Specialized handling of FAQs: FAQs are a common form of knowledge - and we have enhanced our search retrieval systems to recognize and index FAQ entries to provide customers with exact answers where relevant.
What’s Next?
We're continuously innovating to enhance our offerings. Soon, we'll be rolling out:
- Support for additional data sources like Google Docs, and Intercom KBs
- Users should be able to index and find answers from existing Slack channels soon.
- In addition - we are also working on making our Search and Automated-Answers accessible through additional interfaces - like APIs, Slack Workflow Steps, and Emoji-based manual triggers.
Wrapping Up
The introduction of the Virtual Agent by ClearFeed represents a leap forward in automated customer support on Slack. By combining instant response capabilities, deep knowledge base integration, and highly configurable workflows, ClearFeed empowers teams to deliver superior support experiences at scale.
If you’d like to learn more on how the Virtual Agent can elevate support on Slack for your organisation, you can reach out to us at support@clearfeed.ai or book a free demo with a support specialist here. You can also join our community Slack channel to try out the product for free.
