Advanced

The Advanced page allows you to configure key parameters of your AI agent, including selecting a language model, setting prediction confidence, and defining inactivity timeouts.

1. Select Your Own LLM

This section enables you to choose an integrated Language Model (LLM) to power your AI agent.

• LLM Provider: Select an available provider (e.g., Moveo.AI, OpenAI, Azure Open AI, Anthropic, Gemini, DeepSeek).
• Model: Once you've selected a provider, choose one of its specific models that best fits your needs. For example, if you select OpenAI, you might choose a model like o1-preview or gpt-4-turbo; if you choose Anthropic, you might opt for claude-3-5-sonnet-20240620.

2. Prediction Threshold

This setting controls the confidence level required for an intent prediction to be considered correct. Adjust the threshold between 0 and 1.

• Lower threshold → More flexible, but may increase false positives.

  In a customer support conversation, a lower threshold lets the AI Agent to be more flexible. Even if it’s not entirely certain about a user’s intent, it will still attempt a relevant response. This helps keep conversations flowing, but can sometimes lead to more incorrect predictions.

• Higher threshold → More conservative, requiring clearer user input.

  For a legal or financial assistance, a higher threshold ensures that only the most confident predictions are accepted. This approach reduces the risk of errors, but may require users to phrase their questions more precisely for the AI Agent to respond accurately.

3. Inactivity Timeout

Defines the duration before the AI forgets the context of an ongoing conversation due to user inactivity.

• Available options: 5 min, 15 min, 30 min, 45 min, 1 hour, 2 hours, 8 hours, 24 hours.

Adjusting these settings optimizes the AI agent's performance based on your specific workflow requirements.

• Short Inactivity Timeouts (e.g., 5–30 min): are ideal in scenarios where quick, continuous interaction is critical. Short inactivity timeouts help the AI agent discard outdated or irrelevant context. This prevents the agent from mistakenly using stale information that could confuse or mislead the ongoing conversation
• Longer Inactivity Timeouts (e.g., 2, 8, 24 hours): are beneficial when conversations are expected to span longer periods. For instance, during complex troubleshooting sessions or extended customer support cases, a longer timeout allows the AI Agent to maintain context even if the conversation pauses for several hours, ensuring continuity when the user returns.