> For the complete documentation index, see [llms.txt](https://gdplabs.gitbook.io/glchat/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://gdplabs.gitbook.io/glchat/application-ui/how-to-guide/chat-configuration/smart-mode.md). # Smart Mode #### Overview Smart Mode helps users interact with GLChat more easily by allowing them to ask questions without manually choosing a specific mode first. Instead of selecting whether to use LLM, Search, Deep Research, AI Agent, or Chat to Connector, users can simply type their prompt. GLChat will automatically choose the most suitable mode based on the request and the available configuration. This is useful for users who want a faster and simpler experience, especially when they are unsure which mode should be used.

Smart Mode On

*** #### How Smart Mode Works When Smart Mode is enabled, GLChat analyzes the user’s prompt and automatically decides the best way to respond. For example: * If the prompt is a general writing or brainstorming request, GLChat may use **LLM Mode**. * If the prompt requires recent or external information, GLChat may use **Search**. * If the prompt requires broader research and comparison, GLChat may use **Deep Research**. * If the prompt needs to interact with an available connector, GLChat may use **Chat to Connector**. * If the prompt requires a more complex workflow, GLChat may use an **AI Agent**. With Smart Mode, users do not need to understand the differences between each mode before asking a question. GLChat handles the mode selection automatically. *** #### Key Benefits **Ask First, Choose Less** Users can focus on writing their question or instruction instead of deciding which mode to use. For example, instead of manually selecting Search or Deep Research, the user can simply ask: > “Find the latest regulation update and summarize the key points.” GLChat will determine the most appropriate mode based on the prompt. *** **Automatic Mode Selection** Smart Mode can route the prompt to the most relevant capability available in the project, such as: * **LLM Mode** for general answers, drafting, rewriting, or brainstorming. * **Search** for current or external information. * **Deep Research** for deeper research, comparison, or multi-source synthesis. * **AI Agent** for internal workflows or more complex tasks. * **Chat to Connector** for direct connector-based actions or queries. The available options depend on the project and organization configuration. *** **Uses the System Default LLM for LLM Mode** When Smart Mode routes a prompt to LLM Mode, GLChat will use the system default LLM configured for the project. The system default LLM follows the first LLM shown in the LLM selection list. This means users do not need to manually select a model when using Smart Mode. The response will follow the default LLM configuration that has already been set for the project. *** **Smart Mode Configuration is Managed by Admin** Smart Mode behavior is controlled from the Admin Dashboard. Admins can configure whether Smart Mode is available and define which capabilities can be used, such as: * Whether Smart Mode is enabled or disabled. * Which modes are allowed. * Which AI Agents are eligible to be selected. * Which connectors are available. * Whether certain capabilities are restricted based on user role or organization policy. This allows organizations to provide a simpler user experience while still maintaining control over what GLChat is allowed to access or execute. *** #### Manual Mode is Still Available Smart Mode is designed to reduce upfront decisions, not remove user control. Users can still switch to a specific mode manually when needed. For example, if a user specifically wants to use Search, Deep Research, or a certain Agent, they can still choose it directly. This gives users flexibility: they can rely on Smart Mode for convenience, or manually choose a mode when they already know what they want.

Smart Mode Off

*** #### Example Usage
User PromptPossible Smart Mode Selection
“Rewrite this message to be more concise.”LLM Mode
“What is the latest update from Bank Indonesia?”Search
“Compare AWS, GCP, and Azure for enterprise deployment.”Deep Research
“Create a GitHub issue for this login bug.”Chat to Connector
“Analyze our support tickets and summarize the top issues.”AI Agent
The final selected mode may depend on the project setup, available connectors, available agents, and admin configuration. *** #### Visibility in the Response When Smart Mode answers a prompt, GLChat can show which mode was used to generate the response. This helps users understand how GLChat handled their request and gives better transparency when the system automatically chooses a mode. For example: > Answered by: Smart Mode → Search or > Answered by: Smart Mode → AI Agent *** #### When to Use Smart Mode Use Smart Mode when: * You are not sure which mode to choose. * You want to ask quickly without configuring the mode first. * Your prompt may require different GLChat capabilities. * You want GLChat to automatically decide the best available path. For more specific cases, you can still manually select a mode anytime. *** #### Summary Smart Mode makes GLChat easier to use by allowing users to type their prompt first and let the system choose the most suitable mode automatically. It reduces the need to manually decide between LLM, Search, Deep Research, AI Agent, or Chat to Connector, while still allowing manual override when needed. Smart Mode also follows the configuration set in the Admin Dashboard, so organizations can keep control over available modes, agents, connectors, and policies.