Shape Their Persona

Each employee has an identity, personality, and brain you can customize. These settings control who the employee is, how they communicate, and how capable they are.

TL;DR

Go to Workspace, select an employee, open the Profile tab. Three sections: Identity (name, role, avatar), Persona (personality and communication style), and Brain (AI model, thinking depth, max actions per turn). Every employee comes with defaults from their role template. Changes save automatically when you click away from a field.

What Is on the Profile Tab

The Profile tab is divided into three sections:

Identity

Controls who the employee is.

Field	What It Does
Avatar	The employee's profile picture (customization coming soon)
Name	Display name used in the sidebar, chat, and across the platform
Role	Job title shown on the employee's profile (e.g., "Content Writer," "Marketing Strategist")

Persona

Controls how the employee communicates.

Field	What It Does
Persona Prompt	A text field where you describe the employee's personality, tone of voice, and communication style. Write naturally: "You are friendly and concise. You use simple language and avoid jargon."

The persona prompt shapes the style of every response. A formal persona produces structured, professional output. A casual persona produces conversational, relaxed output. The persona does not affect the quality of the work, only the style.

Brain

Controls the AI model powering the employee and how deeply it thinks.

Field	What It Does
Model	Which AI model the employee uses. Dropdown shows all available models grouped by tier (Fast, Standard, Advanced)
Thinking Depth	How much the model "thinks" before responding. Only active for models that support extended reasoning
Max Actions per Turn	How many actions (tool calls, web searches, file writes) the employee can take in a single response. Leave empty to use the organization default

AI Model Tiers

Models are grouped into three tiers that balance speed, capability, and cost:

Tier	Best For	Credit Cost
Fast	Quick tasks like drafts, lookups, and formatting	Lowest (60 to 303 credits per 1M tokens)
Standard	Most everyday work like writing, research, and planning	Medium (705 to 1,603 credits per 1M tokens)
Advanced	Complex reasoning, strategy, analysis, and coding	Highest (2,120 to 3,915 credits per 1M tokens)

The model dropdown shows models from multiple providers (OpenAI, Anthropic, and others). You can also select local models via Ollama, which use zero API credits.

Supported Languages

Strongest in English first.

Each model has different capabilities, but you can always find your language in one of our supported languages.

Albanian, Amharic, Arabic, Armenian, Bengali, Bosnian, Bulgarian, Burmese, Catalan, Chinese, Croatian, Czech, Danish, Dutch, Estonian, Finnish, French, Georgian, German, Greek, Gujarati, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Marathi, Mongolian, Norwegian, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Slovak, Slovenian, Somali, Spanish, Swahili, Swedish, Tagalog, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Vietnamese and much more.

How Model Selection Works

The system resolves which model to use in this order:

Employee override (set on the Profile tab) takes priority
If no override is set, the workspace default is used (set in Settings, under AI Behavior)
If no workspace default is set, the system default is used

To set the same model for all employees at once, click the "Change for all employees" link in the Brain section. This takes you to the AI Behavior settings where you can pick a workspace-wide default.

Thinking Depth

For models that support extended thinking, you can control how deeply the employee reasons before responding:

Level	What It Does
Default	Uses whatever the system or organization default is
Auto (Recommended)	Adjusts dynamically. Quick answers for simple messages, deeper thinking for complex ones
Off	No reasoning. Fastest responses, lowest cost
Minimal	Lightest reasoning
Low	Light reasoning
Medium	Moderate reasoning
High	Maximum reasoning depth. Best for complex analysis, but uses more credits

The thinking depth dropdown is always visible. If the selected model does not support reasoning, the dropdown is disabled (grayed out).

What You Can Do

Action	How
Change the name	Profile tab, Identity section. Edit the Name field, click away to save
Change the role	Profile tab, Identity section. Edit the Role field, click away to save
Set personality	Profile tab, Persona section. Write a persona prompt describing tone and style
Switch AI model	Profile tab, Brain section. Select from the Model dropdown
Set thinking depth	Profile tab, Brain section. Select from the Thinking Depth dropdown
Set max actions	Profile tab, Brain section. Enter a number (50 to 1000) or leave empty for org default
Change model for all employees	Profile tab, Brain section. Click "Change for all employees" link
Use a free local model	Profile tab, Brain section. Select an Ollama model (requires Ollama running locally)

How to Set It Up

Go to Workspace, click on an employee in the sidebar
Open the Profile tab
Edit any field. Changes save automatically when you click away

Tips and Tricks

Match the persona to the role. A customer-facing writer might be friendly and conversational. A legal analyst might be formal and precise.
Use Fast for simple tasks. Quick lookups, formatting, and drafts cost significantly less than Standard or Advanced models.
Set different models per employee. Your content writer can use Standard while your research analyst uses Advanced. Each employee's model is independent.
Persona is not a skill. Do not put process instructions in the persona prompt. Use the Skills tab for workflows and the Duties tab for standing rules.
Try Auto reasoning. The "Auto" thinking depth is the best default for most use cases. It adapts based on the complexity of each message.

Good to Know

Changes apply immediately. Updated persona, model, and reasoning settings take effect on the next message. No restart needed.
No quality impact from persona. Persona changes style, not capability. A casual employee produces the same quality work as a formal one.
Per-employee model. Changing one employee's model does not affect any other employee.
Provider fallback. If a model provider goes down (e.g., OpenAI has an outage), the system automatically retries with a model from another provider. You are billed at the fallback model's rate.
Ollama is free. Local models via Ollama use zero API credits. Useful for testing or when you want full privacy.
Auto-save on blur. All fields save automatically when you click outside (blur). There is no save button.

Frequently Asked Questions

Q: Does changing the personality affect the quality of work? A: No. Persona changes the style of communication (formal vs. casual, detailed vs. concise) but not the quality or accuracy of the output.

Q: Can I use AI models from different providers? A: Yes. The Model dropdown includes models from OpenAI, Anthropic, and Ollama (local). Select any model that fits your needs and budget.

Q: What happens if the AI provider goes down? A: The system automatically retries with a fallback model from a different provider. You are billed at the fallback model's rate, not the original.

Q: Does using Advanced always produce better results? A: For complex reasoning, strategy, and coding, yes. For simple tasks like formatting or quick lookups, Fast models are equally effective and cost significantly less.

Q: Can I change the employee's name? A: Yes. Edit the Name field in the Identity section of the Profile tab. The employee responds to whatever name you give them. The change takes effect immediately.

Q: What does "Max Actions per Turn" mean? A: It limits how many actions (tool calls, web searches, file operations) the employee can perform in a single response. Higher values allow longer, multi-step work. Leave it empty to use your organization's default (typically 250).