Introduction

Khoros Generative Models utilize large language models (LLMs) and customer-provided knowledge to generate automated responses. Designed to enhance customer care and community management, Khoros Generative Models simplifies how organizations handle customer interactions, making customer service more efficient and responsive. By leveraging AI-driven automation, businesses can respond to inquiries faster, provide accurate information, and improve customer satisfaction.

What is the Khoros Generative Models?

Khoros Generative Models is the generative AI engine powering Khoros' customer service and community engagement platforms. Unlike traditional chatbots that rely heavily on pre-defined intents and training, Khoros Generative Models use LLMs to create dynamic, context-aware responses. These responses can be generated from various knowledge sources, including question-answer pairs, customer websites, or community pages. This capability allows businesses to offer more personalized and accurate answers to customer inquiries without needing extensive manual training and flow designing for the bot.

Example:

Consider a customer querying a technical issue on a website. Traditional chatbots might struggle if the inquiry does not match pre-defined intents. However, Khoros Generative Models can dynamically pull relevant information from a company's knowledge base, community threads, or FAQs to provide a precise answer.

Core Components of Khoros Generative Models

Khoros Generative Models is built upon several key components that work together to deliver a seamless customer experience:

• Flow Messaging Project: This is the conversation design platform where chatbot flows are created, and intents are trained manually. It provides a foundation for integrating Khoros Generative Models’s capabilities with other chatbot features.
• Knowledge Source: This component contains the knowledge the generative models will use. It can consist of question-answer pairs, Community pages, other web pages, terminology, or a mix of multiple different sources.
• Persona: This is a collection of configuration settings, including the prompt template, knowledge sources, and LLM settings. The persona defines how Khoros Generative Models responds, whether as a chatbot or for agent assist.

Detailed Explanation:

The Flow Messaging Project allows for the manual creation of chatbot flows, enabling businesses to customize responses based on specific user inputs. Meanwhile, Knowledge Sources are vital as they contain the information Khoros Generative Models will reference. Khoros Generative Models Persona ensures that responses are tailored to the specific context of the conversation, allowing for variations in tone and depth depending on the situation.

Setting Up Khoros Generative Models

1. Add a Khoros Generative Models Knowledge Source

   1. Begin by creating a new Knowledge project to manage your knowledge sources.
   2. Click the "+" button to add a new knowledge source. You can input question-and-answer pairs directly into the UI or upload them via Excel or CSV files.
   3. You can add multiple knowledge sources tailored to distinct subject areas if your business operates across different business units.

2. Configure a Persona

   1. Navigate to Organization Settings > Personas and click the "+" button to add a new persona.
   2. Choose the persona type (bot or agent assist) and select a prompt template. Give the persona a descriptive name and save it.
   3. Optionally, add a description.
   4. Select which knowledge sources the persona should have access to.

3. Connect a Flow Bot with Khoros Generative Models

   1. In your Flow bot project, go to Configuration > Languages. Select the desired language and enable Generative AI.
   2. Choose between the Global or Custom configuration types.
   3. Select the appropriate persona for the bot project.
   4. If you use the Custom type, add one or more Dynamic Replies to your bot Flows and configure them in the right panel.

How Khoros Generative Models Works

Khoros Generative Models can operate in two primary modes: Global and Custom configurations. The Global configuration allows Khoros Generative Models to take control automatically when no matching intent is found, ensuring a smooth conversation flow without interruption. In contrast, the Custom configuration provides more granular control over when Khoros Generative Models intervenes, allowing businesses to fine-tune the chatbot's responses.

Scenario:

For example, in a Global configuration, if a user asks a question that doesn’t match any pre-configured intents, Khoros Generative Models will immediately search their knowledge base and generate an appropriate response. Conversely, in a Custom configuration, the chatbot could be set to first attempt to match the intent and then decide whether to use Khoros Generative Models based on the context of the conversation if the bot design leads to a Dynamic Reply.

Leveraging Khoros Generative Models in Customer Care

Khoros Generative Models enhance the Customer Care Cloud (CCC) by enabling more efficient responses and providing conversational insights. It serves as a conversational search tool that retrieves and summarizes knowledge from the community, offering answers to users' questions. If Khoros Generative Models cannot find a suitable answer, it hands off the conversation to a human agent and uses the agent's response to draft new articles for future reference.

Benefits in Customer Care:

• Reduced Manual Effort: With Khoros Generative Models, less manual building and training are required, enabling conversation designers / bot designers to focus on more complex issues.
• Improved Response Time: Khoros Generative Models listens to conversations and suggests relevant responses, reducing the time agents spend searching for answers.
• Enhanced Tagging and Summarization: Khoros Generative Models automatically creates a taxonomy and tags conversations, making it easier for agents to track and manage issues. Summarizations provide agents with a concise conversation overview, including user sentiment and needs.

Using Khoros Generative Models in Community Management

Khoros Generative Models also plays a vital role in community management by searching community articles and threads, summarizing answers, and citing sources. This functionality helps community managers maintain an up-to-date knowledge base and ensures that common questions are addressed proactively.

Use Case:

A community manager may notice that users frequently ask the same questions. Khoros Generative Models can analyze these questions, generate summarized responses from existing content, and draft new articles to address recurring queries. This proactive approach not only improves the community experience but also reduces the workload for managers.

Benefits of Using Khoros Generative Models

Khoros Generative Models offers numerous advantages, making it an invaluable solution for businesses. Some of the key benefits include:

• Increased Efficiency: Khoros Generative Models reduces the number of manual tasks involved in chatbot training and customer service, allowing conversation designers and customer care agents to focus on higher-value activities.
• Improved Customer Satisfaction: By providing accurate and timely responses, Khoros Generative Models enhances the customer experience, leading to higher satisfaction rates.
• Better Conversation Management: With features like automatic tagging, summarization, and sentiment analysis, Khoros Generative Models ensures that conversations are well-organized and easy to manage.
• Multi-Intent Handling: Unlike traditional chatbots, Khoros Generative Models can understand and respond to multiple intents in a single query, reducing confusion and improving the relevance of responses.

Advanced Features: AI Insights and Summarizations

Khoros Generative Models includes advanced features like AI Insights and Summarizations, further enhancing its capabilities.

• AI Insights: These are automatically generated system tags that categorize conversations based on content. They provide valuable data for analytics, helping businesses understand common issues and areas for improvement.
• Summarizations: The Summarization API allows Khoros Generative Models to create concise recaps of bot-user conversations, highlighting key details such as user sentiment and next steps. This feature is particularly useful during handoffs to human agents, enabling quicker and more informed responses.

Conclusion

Khoros Generative Models represent a significant advancement in how businesses interact with customers and manage communities. By leveraging the power of generative AI, Khoros Generative Models automates responses, streamlines customer service, and improves the overall user experience. As organizations continue to adopt AI-driven technologies, tools like Khoros Generative Models will become increasingly essential in maintaining a competitive edge.