Conversation Graph

Conversation Graph transforms your historical support tickets into a structured knowledge base, enabling Siena to provide consistent, accurate responses based on how your best agents have already solved similar problems. Your customers benefit from solutions proven to work, while your team gains higher automation rates without manual knowledge base creation.

How It Works

Conversation Graph automatically processes your support history through a sophisticated pipeline. There are five key process steps:

Siena will automatically filter and process only high-quality conversations that contain valuable knowledge, excluding tickets that are too brief, too verbose, spam, or contain only bot responses.

Some examples of what can be captured as knowledge:

• Policy exceptions and edge cases your team has handled
• Step-by-step troubleshooting processes
• Product-specific solutions and workarounds
• Escalation procedures and decision-making criteria

To start using Conversation Graph, navigate to the Knowledge Sources page within Siena Settings, then connect your helpdesk platform (Gladly, Gorgias, Kustomer, Zendesk). The feature will automatically begin extracting and processing your past 14 days of support ticket data.

You can then configure the Look Back period - this determines how far back in time into your historical conversations Siena should look, using their created at date as a marker, when attempting to use them as a Knowledge Source. Choosing 180 days would begin eliminating conversations from eligibility once they are over 180 days old.

Once the extraction is complete you'll be able to see which conversations are being used to generate responses in real-time.

How Conversation Graph Knowledge Is Used

This knowledge is seamlessly integrated into Siena's responses, allowing it to provide solutions based on your team's actual problem-solving approaches. Siena can leverage this knowledge to:

• Automatically handle routine questions with consistent answers
• Apply policy-based reasoning from previous similar cases
• Reference successful solutions your agents have used before
• Understand edge cases and exceptions based on past tickets

What You'll See

In your conversations, you'll notice Siena handling more inquiries autonomously and providing responses that match how your best agents would solve problems. For example, if a customer asks about a return policy exception that your team has handled before, Siena will automatically apply the same reasoning and solution.

You'll also be able to see in the Knowledge Sources table when conversations from your Conversation Graph are being used, and explore the details of specific conversations that informed Siena's responses.

Clicking an individual row for conversation details, you can see exactly which customer interactions benefited from this historical knowledge and maintain full traceability.

When reviewing your conversations in the Inbox, you'll be able to see these historical conversations inline in the Knowledge Sources pop-over.

Analyzing Your Conversation Graph

Your Conversation Graph knowledge refreshes automatically every 24 hours, continuously learning from new support interactions. You can monitor automation rates and customer satisfaction scores to ensure the knowledge is improving your support quality and efficiency over time.

The system preserves institutional knowledge when agents leave and gets smarter with every ticket processed, creating a continuously improving foundation for automated customer support.