> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.ada.cx/docs/setup/persona/best-practices/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ada.cx/_mcp/server. # Best practices ## Overview Your AI Agent's Persona is more than just a setting: it is how your brand comes to life in every interaction. A clear, well-defined Persona helps build trust, keeps your brand voice consistent, and ensures end users feel like they are engaging with a reliable representative of your business. A well-configured Persona: * Builds end user trust by aligning the AI Agent with your brand voice * Reduces confusion with consistent tone across channels * Improves engagement by reflecting the qualities of your best representative ## Configure identity Set the identity details that shape how your AI Agent shows up to end users: * **Agent name**: Use a human, brand-attached name (e.g., *Ada from Acme*) so messages feel personal and trustworthy. Avoid generic or technical labels like *Assistant*, *noreply*, or internal codes. * **Company name**: Match the public-facing brand exactly (spelling/capitalization as on your site), not the legal entity name, to keep recognition high across channels. * **Company description**: Keep it to one clear sentence that explains what you do (e.g., *Acme makes eco-friendly outdoor gear!*). Avoid slogans or vague taglines. * **Default time zone**: Set the zone end users expect for time references (e.g., *\[UTC-04:00] America/Toronto - EDT*). ### Example: Identity settings (good vs. not-so-good) Use these examples to configure the AI Agent in a way that feels human, on-brand, and clear for end users. | Field | ✅ Good example | ❌ Not so good | | ----------------------- | ------------------------------------- | ------------------------------- | | **Agent name** | Ada from Acme | Ada Bot, bot123 | | **Company name** | Acme | ACME INC. LTD., Acme Internal | | **Company description** | Acme makes eco-friendly outdoor gear. | We innovate synergy.; The best! | ### Email-specific settings * Ensure the **Sender name** found on the **[Customization tab in your Email settings](/docs/channels/email/email-configuration/appearance-settings)** matches what end users should see in email threads. * The AI Agent signs off using the **Agent name** field. #### Example: Identity settings for Email | Example | Sender name | Agent name | | ------- | ------------ | --------------------- | | ✅ | Acme Support | Thanks, Ada from Acme | | ❌ | bot123 | Ada Bot | ## Define tone Work with your team to decide on your AI Agent's qualities, tone, and style. In the **Tone and style** tab, under **General**, set **Tone of voice** to **Friendly**, **Plainspoken**, **Playful**, or **Sophisticated**. ### Example: Configure tone If your brand is approachable and consumer-facing, and you want empathetic, low-friction support for common issues (e.g., refunds, shipping delays, account hiccups), choose **Friendly**. Here's what your AI Agent's reply may look like when you set the **Tone of voice** to **Friendly**: `Hey Maya! So sorry your order's running late. I've checked your tracking and it's now due tomorrow. I've also refunded the shipping fee. Anything else I can help with today?` > **Note** > > The sample above is an illustration. Actual responses depend on the user’s input and your configuration. To verify the effect of **Tone of voice**, set it to **Friendly** and use **Preview** in the Dashboard. If needed, adjust the tone or other Persona settings and preview again. ## Adjust message length Use **Tone and style** settings to control how your AI Agent writes. Review **Message length** (**Normal** or **Concise**, as applicable). Before rollout, use **Preview** in the Dashboard to confirm everything reads as expected. ### Examples: Outputs from Message length settings Here are some sample outputs produced by the **Message length** settings. They illustrate what end users might see when different message length options are applied. | Setting | ✅ Output when configured well | ❌ Output when configured poorly | | -------------------------- | -------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | **Message length (Chat)** | `To reset your password, go to Settings > Password > Reset.` | `Thanks for contacting Acme! Please update the app and check your connection, then go to Settings > Password > Reset, request a code, check spam, wait 5–10 min, enter the code, and create a compliant password.` | | **Message length (Email)** | `Hi Alex — here are the steps: 1. Settings > Password > Reset. 2. Send code. 3. Enter code.` | `We appreciate your valued business. To reset your password you will need to navigate to the application settings where you will find the password section and then proceed accordingly.` | ## Consider brand and context * **Brand alignment:** Model the persona on your ideal representative to build trust. > **Tip** > > **Example**: A fintech company might choose a **Plainspoken** or **Sophisticated** tone, similar to how a trusted financial advisor speaks, while a retail brand could adopt a **Friendly** or **Playful** style that mirrors in-store staff. * **Multilingual consistency:** If you [support multiple languages](/docs/setup/languages/about-multilingual-support), review tone per locale and use language customization options to maintain consistency. > **Tip** > > **Example**: In English, the AI Agent may sound light and playful (e.g., `Got it! Here's a quick fix!`), but in Japanese the same response may be phrased more formally to match cultural expectations. * **Ongoing review:** Revisit your Persona settings regularly as brand and business needs evolve. > **Tip** > > **Example**: A startup might begin with a **Playful** tone to reflect its early culture, but later adjust to a **Sophisticated** tone as it moves into enterprise markets. * **Channel-specific settings:** Some Persona options only apply to certain channels. * **Chat/Social/Email:** Includes tone and style controls like **emoji usage**. > **Tip** > > **Example**: `Your order is on the way 🚚` (with emojis enabled) vs. `Your order is on the way` (emojis disabled). * **Email:** Has [unique setup](#email-specific-settings) to specify the **Sender name** and sign-off. > **Tip** > > **Example:** [See here](#example-identity-settings-for-email). ## Do / Don’t **Do** * Align persona with existing brand guidelines * Preview sample responses to confirm tone and style * Keep Persona settings updated as your brand evolves **Don't** * Overlook Email-specific settings like **Sender name** and sign-off * Assume tone translates identically across all languages without review ## Final checklist Before finishing setup, make sure you've completed all of the following tasks: | ✅ Task | | ------------------------------------------------------------------------------- | | Defined the AI Agent's persona qualities, tone, and style | | Configured identity settings (Agent name, Company name, Description, Time zone) | | Adjusted Tone and Message length | | Set Email-specific fields (Sender name, Agent name) | | Previewed responses in the Dashboard to confirm alignment | | Planned for multilingual consistency (if applicable) | | Scheduled periodic reviews to update the Persona over time | --- Have any questions? Contact your Ada team, or email us at [](mailto:help@ada.cx?subject=Help%20Docs%20inquiry).