Solve Widget | Deflection Insights Guide
We recently released a guide on our latest feature, Deflection Insights. This new feature helps improve customer support by evaluating the quality of interactions between users and the chatbot. Unlike traditional Deflections, which only measure whether a conversation was redirected from human agents, Deflection Insights focuses on the quality of chatbot responses and user engagement. To learn more about Deflection Insights, check out this article.
Solve Widget | Announcing the general availability of Solve Insights Article
The Solve Insights Article feature is now available for all customers! This feature enables customers to analyze their article performance comprehensively and access various metrics. For more information, see this article.
Analytics | Filter conversations by any Context Variables (CV) value in the Solve Chat Analytics looker dashboard
We are excited to announce a small yet impactful improvement to our Looker analytics page! With this new update, you can now filter conversations by any CV value. This means that you can now easily analyze the performance of your chatbot across different CV values. To access this filter, simply navigate to your Forethought dashboard > Analytics > Solve Chat Analytics.
Solve Widget | Chatbot responses that leverage your existing Knowledge Base are now more human-like
Starting Apr 18, all Widget customers will see an automatic improvement to the chatbot's response when paraphrasing information collected from your connected Knowledge Base. The response will look more human-like, have neater formatting, retain memory of the previous conversations, and even ask clarifying questions in case users ask vague questions. For more information, see this article.
Assist | Improvement in Assist Feed and Response Generation Performance
No action is required from you to enjoy this update. Today, we are enhancing the ranking performance for all Assist customers. Our updated backend ranking algorithm will significantly improve accuracy compared to current capabilities. The Feed will now prioritize clean and precise articles, placing them at the top, while tickets, notes, and macros will be positioned below these articles. For generating initial responses, we will exclusively use the most relevant articles from the Feed. Agents still have the option to include free-form text, which allows for the inclusion of information from text, notes, or macros, maintaining flexibility while greatly enhancing both the Feed and initial response generation.
Solve Widget | Capturing every intent for deeper insights in the landing page metrics
We improved the landing page metrics to understand user interaction with your chatbot better. Previously, the metrics only considered the last intent used, resulting in data loss if there were multiple intents. To address this, we changed the way we calculate the chat column logic. This article explains the metrics and demonstrates the new chat column logic with examples.
Solve Widget | Identify gaps in your knowledge base with Fallback
When a chatbot cannot find relevant information in its knowledge base, it activates Fallback. Fallback helps identify areas where more knowledge base articles are needed, thereby improving support for end-users. If Knowledge Retrieval is disabled, Fallback will activate whenever a custom Intent does not match, providing a method to engage with the user before a handoff occurs. Refer to the article for more information and an example of how Fallback works.
Solve Widget | Personalize Your Customer Satisfaction Survey (CSAT) to Fit Your Needs
Customize your Customer Satisfaction Survey (CSAT) to fit your requirements and preferences. Access the customization feature directly within the Widget to personalize your CSAT. Gather more relevant insights from your customers to improve your business and provide a better customer experience. Read more about this feature in the article.