Optimize Knowledge Content for Large Language Models
Learning Objectives
After completing this unit, you’ll be able to:
- Describe grounding in generative AI.
- Explain writing and process guidelines for knowledge bases.
What Is AI Grounding?
Grounding means connecting your AI model to trusted sources of information. Because AI systems use machine learning and large language models to create new content, they’re only as good as the data on which they’ve been trained. If this seems risky, the good news is that your knowledge base is probably a great source of trusted information on which to train your AI model. Grounding, also known as retrieval-augmented generation (RAG), ensures that your AI-generated content is accurate and specific to your organization.
In customer service, accuracy is everything. Give a customer the wrong answer, and it could discourage them from future purchases. Provide agents with incorrect information, and they might frustrate customers rather than support them. By grounding AI in verified sources like your knowledge base, generative AI can create responses to customer inquiries, suggest replies to agents, and provide search summaries—all with accuracy. But to ensure accuracy, you must prepare your knowledge base for large language models.
Prepare Your Knowledge Base
Your knowledge base in Salesforce is the basis for your organization’s verified information. If you have additional knowledge outside Salesforce—such as in Google Docs, Microsoft SharePoint, Atlassian Confluence, or elsewhere—you can import and sync it with Unified Knowledge.
Unified Knowledge integrates all your third-party knowledge in Salesforce. Once all your knowledge articles are in Salesforce, agents and customers can find the information they need from one platform. Additionally, Unified Knowledge uses your various knowledge sources to ground Agentforce for Service and its generative AI capabilities.
Grounding all your knowledge improves the usefulness of Service Replies—AI-generated suggested responses to customer inquiries. Similarly, this grounding enhances the effectiveness of Search Answers and Agentforce Service Agent—the Salesforce tools for AI-generated search summaries and AI assistance for service agents and customers.
To ensure your AI-generated content is precise and tailored to your organization, follow these simple writing and process guidelines.
Knowledge Base Writing Guidelines
To improve grounding results, consider these guidelines for writing knowledge articles.
Explain Thoroughly
Unlike humans who have comparatively short attention spans, generative AI works best with detailed, comprehensive information. It synthesizes this data to deliver insights at various levels of detail for users. Always lean toward including more information, especially technical details.
Provide Examples
Think about common, real-world situations users face and write about them with specific examples. Describe examples in a conversational way to help AI interpret relevance to users.
Structure Content
AI, like humans, prefers structured content. Write sentences with logical relationships and break up content into paragraphs or lists. Use heading tags (H1 to H6) for a hierarchical structure, and separate internal and customer-facing information into different fields.
Annotate Media
If you use visual content in articles, such as videos, screenshots, or animated gifs, annotate them with textual descriptions. Simple images are great for humans, but AI requires captions and alt-text to understand them.
Answer Common Questions
Most users interact with generative AI in the form of questions. If you write FAQs or answers to specific questions, AI generates more helpful responses.
Create Focused Content
Write articles that cover a single topic instead of including 10 issues in one article. Focused writing helps you manage security and visibility, as certain topics are appropriate for some audiences but not others. Also, focused content helps AI generate more relevant output.
Next, to ensure that your knowledge base remains a reliable and effective resource for generating AI output, consider the processes needed to manage it.
Knowledge Base Process Guidelines
To improve grounding results, consider these process guidelines for knowledge articles.
Follow KCS Principles
Knowledge-Centered Service (KCS) is a set of industry-standard best practices on knowledge management. KCS helps you create accurate and consistent articles, which improves knowledge grounding. To learn more, check out the Consortium for Service Innovation website.
Audit Your Knowledge Base
If you don’t regularly review and update articles, your AI-generated content will include incorrect and outdated information. Use Salesforce Flows to build logic on articles’ Next Review Date field to identify content for review.
Remember Humans
Balance managing content that’s consumable by AI but useful and understandable to humans. To build trust and help users dive deeper into information, add citations and links to what your knowledge is grounded in. References to information gives credibility to generated content.
Conclusion
Now that you’ve learned about AI grounding, along with some writing and knowledge management processes, you’re ready to ensure that AI-generated content from Agentforce for Service is accurate and specific to your organization. Get started today by organizing an audit of your knowledge base, and use Unified Knowledge to bring in your trusted content from other sources!