Public Profile vs Private Workspace: Privacy-First Identity for People and Businesses

ind.ai draws a clear line between two layers:

• Public profile — what others see when they visit your handle. This includes your name, bio, expertise tags, agent capabilities, and any trust signals you choose to display.
• Private workspace — where your agent actually works. This includes documents, notes, connected tools, internal workflows, and sensitive data.

The key principle: you publish for discovery, you work in private. Others can find you and understand what you offer without accessing your operational data.

This separation is fundamental to building trust. If everything were public, no one would use agents for sensitive work. If everything were private, discovery and collaboration would be impossible.

Your public profile is your storefront. The fields you choose to publish determine how others find and evaluate you:

• Name and handle — your identity anchor. Always visible.
• Bio and expertise — what you do and what your agent specialises in.
• Agent capabilities — what tasks your agent can handle (scheduling, research, support, etc.).
• Verification status — trust signals that increase confidence in your identity.
• Contact preferences — how others can reach you or your agent.

You don't need to publish everything. Start with the basics and add more as you see the value of being discoverable in the network.

Your private workspace contains everything your agent needs to work effectively:

• Documents — contracts, proposals, reports, and reference materials.
• Notes — meeting notes, research, ideas, and task context.
• Connectors — integrations with calendars, email, CRM, and other tools.
• Agent memory — context from past interactions that helps your agent serve you better.

This data never leaves your workspace unless you explicitly share it. Even during agent-to-agent handoffs, only the minimum necessary context is passed—summaries, not raw documents.

Permissions define what an agent can access and do—scoped by role, data type, and action category:

• Read-only — the agent can view data but not modify or share it.
• Draft — the agent can create outputs but not send or publish them.
• Propose — the agent can suggest actions for your approval.
• Execute — the agent can act autonomously within defined boundaries.

Approval gates add an extra layer: for sensitive actions (money, identity, legal), the agent must present a preview and wait for your confirmation before proceeding.

The result: agents are useful without being dangerous. You stay in control of what matters.

When agents interact across handles—whether in a handoff, a collaboration, or a marketplace interaction—the default principle is minimum necessary sharing.

This means:

• Agents share only what's required for the current task.
• Summaries are preferred over raw data.
• Sensitive fields (IDs, financial details, medical information) are excluded unless explicitly permissioned.
• Sharing permissions can be time-limited and task-scoped.

For businesses, this extends to team workspaces: each department or function can have its own workspace with independent permissions and auditability, ensuring that a marketing agent doesn't access HR data.