Advanced
The Advanced page contains settings that influence how the agent remembers users, manages conversation memory, handles time-sensitive data, and which AI model it uses for different tasks.
Navigate to Persona → Advanced in the left sidebar.
Personalization
Personalization allows the agent to remember and learn about each user over time, then use that knowledge to provide more relevant, context-aware responses in future conversations.
When a user interacts with the agent, the system automatically extracts personal information from the conversation — things the user mentions about themselves, what they own, and what they prefer. This information is categorized into three types:
| Category | What is captured | Examples |
|---|---|---|
| Identity | Basic profile information | Name, gender, hometown, birthday |
| Ownerships | Things the user owns or manages | "I have a Honda Civic", "I rent an apartment in New York" |
| Preferences | Likes, dislikes, and habits | "I prefer concise answers", "I usually work at night" |
Each piece of information is stored as a semantic embedding in a vector database. On every subsequent conversation, the agent automatically retrieves the most relevant entries — based on the topic being discussed — and incorporates them into its reasoning before responding. This happens in the background, without any action required from the user.
The result is an agent that can refer back to what a user has said in the past, avoid asking repetitive questions, and calibrate its responses to the individual — without the user needing to re-introduce themselves each time.
Personalization data is isolated per agent and per user. One user's profile is never shared with other users or other agents.
Enable Personalization
When enabled, the agent builds a profile of each user over time based on their past interactions. It uses this profile to tailor responses — for example, adjusting the level of detail, remembering stated preferences, or referencing earlier conversations.
Use this for agents where ongoing relationships matter, such as coaches, tutors, or personal assistants.
Leave off for anonymous or transactional agents where user history is not relevant.
Personalization AI Model
The AI model used specifically for reading and updating user profiles. This can be a smaller, more cost-efficient model than the one used for conversation — since profile updates happen in the background and do not require the most capable model.
Timezone Configuration
Agent Default Timezone
Sets the timezone used as the reference point for all date and time reasoning performed by the agent.
This affects:
- Thread separation (conversations are split at midnight in this timezone)
- Any date/time-aware responses or scheduled actions
Example: An agent serving users in Jakarta should be set to Asia/Jakarta (UTC+7).
Choose the timezone that represents your primary user base. Users in other timezones can still interact normally — this setting only affects how the agent interprets "today" and daily boundaries internally.
Conversation Context
These settings control how much of the conversation history the agent "remembers" when composing each response.
Summarize Previous Answer
When enabled, the agent summarizes its own previous responses before including them in the context window. This reduces the number of tokens used to represent prior turns, allowing longer conversations to fit within the model's context limit.
Use this for agents that handle extended multi-turn conversations.
Trade-off: Summaries may lose fine-grained detail from earlier answers. If precision in referencing past responses is important, keep this off.
Keep Function Tools Answer
When enabled, the output of any external tools or function calls (e.g., API responses, database query results) is retained in the conversation context for follow-up turns.
Use this when users are likely to ask follow-up questions about data retrieved by a tool (e.g., "Can you sort that by date?" after a database query).
User Messages to Carry Over
The number of recent user messages to include in the context window when starting a new conversation segment or after thread separation.
| Value | Behaviour |
|---|---|
0 | No history is carried into the new segment |
10 (default) | The last 10 user messages and their responses are included |
| Higher values | More continuity but higher token usage |
Set this based on how much conversational continuity matters for your agent and how long typical conversations are.
Model Capability
Conversation AI Model
The primary AI model used for all user-facing conversation. This model handles every message the user sends and generates the visible response.
Choose based on the trade-off between response quality, speed, and cost:
| Consideration | Guidance |
|---|---|
| High accuracy required | Use the most capable model available |
| High volume / cost-sensitive | Use a smaller, faster model |
| Mixed needs | Enable User to Change AI Model in Dialog Settings |
Enforce Thinking
When enabled, the AI is required to reason through its approach before generating the final response. This improves the quality of answers to complex or multi-step questions.
Use this for agents dealing with analysis, troubleshooting, or any task requiring careful reasoning.
Trade-off: Thinking adds latency. Responses take slightly longer to appear.
Saving Changes
Click Save after adjusting any setting. Publish your agent to propagate the changes to end users.
Related pages
- Behavior — mission and personality
- Dialog Settings — language, formatting, and response suggestions