> ## Documentation Index
> Fetch the complete documentation index at: https://docs.socializioz.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How Socializioz works

> Understand how Socializioz uses workspace context, an AI Agent, direct product surfaces, approvals, and verified execution to run social media work.

Socializioz is built around an operating loop, not a collection of disconnected content tools.

```text theme={null}
Context → Plan → Create → Validate → Review → Execute → Learn
```

The Socializioz Agent and the direct workspace surfaces use the same underlying workspace state. This means a campaign created in chat can appear in Campaigns, a post prepared by an external MCP client can enter the normal approval flow, and a scheduled item remains visible in the Publishing Hub and calendar.

## 1. Context comes first

Before taking action, Socializioz can use the context available to the authorized workspace:

* Active workspace and membership.
* Brand identity, tone, audience, goals, and guidance.
* Connected accounts and their current capabilities.
* Campaigns and content-plan items.
* Existing posts and their current state.
* Media assets available in the workspace.
* Plan access, roles, and approval requirements.
* Available publishing and analytics data.

Good context reduces generic output and prevents the Agent from treating every platform, brand, or campaign as interchangeable.

## 2. The Agent reasons across the workflow

The Agent is not limited to generating text. Depending on the active plan, authorization, and available tools, it can help:

* Plan a campaign from an objective.
* Build and update campaign content plans.
* Create posts from approved ideas.
* Prepare platform-specific variants.
* Find, import, attach, and validate media.
* Route work through approvals.
* Schedule, reschedule, cancel, publish, and retry.
* Inspect current post or publishing state.
* Summarize available analytics and recommend next actions.

The Agent should not claim an action succeeded because it intended to perform it. It should use the returned Socializioz result as the source of truth.

## 3. Direct tools remain available

Conversational execution is optional. You can use the direct workspace surfaces when you want detailed control:

| Surface            | Main purpose                                                            |
| ------------------ | ----------------------------------------------------------------------- |
| Workspace overview | Active brand, channels, operational state, and shortcuts                |
| Campaigns          | Objectives, plans, posts, dates, and campaign performance               |
| Composer           | Detailed content, media, platform settings, and publish controls        |
| Publishing Hub     | Operational status, approvals, failures, retries, and published results |
| Calendar           | Timing, queue distribution, and rescheduling                            |
| Media Library      | Reusable images, videos, imports, generation, and organization          |
| Analytics          | Available performance metrics and insight workflows                     |
| Comments and Inbox | Supported public and private engagement workflows                       |

The Agent and the direct tools are two ways to operate the same system, not two separate products.

## 4. Validation happens before execution

Social platforms have different requirements. Socializioz checks the information available before scheduling or publishing, including:

* Destination account and connection health.
* Required media type.
* Media availability and compatibility.
* Caption and platform constraints.
* Post state.
* Schedule timing.
* Approval state.
* Plan and role permissions.

A failed validation is not a completed publish. Correct the reported problem and retry through the appropriate workflow.

## 5. Human control stays visible

Socializioz can accelerate work without hiding consequential decisions.

Common control points include:

* Reviewing generated content and factual claims.
* Editing the destination, timing, or media.
* Sending work for approval.
* Approving or rejecting a post.
* Choosing whether to publish now or schedule.
* Cancelling scheduled work.
* Retrying a failed publication after reviewing the cause.

The exact approval behavior depends on the workspace, role, plan, and post state.

## 6. Execution is recorded in the workspace

When an action succeeds, the resulting state should be visible in the relevant Socializioz surface:

* Campaign in Campaigns.
* Draft in Composer or Publishing Hub.
* Pending post in Approvals.
* Scheduled item in the calendar.
* Published item in Published Posts.
* Failure and retry state in Publishing Hub.
* Source and action details in Activity Log where available.

This keeps work created by a person, the Socializioz Agent, WhatsApp, ChatGPT, Claude, or another supported MCP client inside one operational system.

## 7. Results inform the next cycle

Analytics is not a separate final tab. Available performance information can help guide:

* Which formats to repeat.
* Which platforms need different content.
* Which campaign ideas performed best.
* Whether timing or consistency changed outcomes.
* What the next campaign should test.

Provider metrics vary, so compare like-for-like data and confirm the active account and date range before relying on a conclusion.

## A complete example

<Steps>
  <Step title="Ask for the outcome">
    `Create a seven-day launch campaign for our new product across Instagram, LinkedIn, and TikTok.`
  </Step>

  <Step title="Use workspace context">
    Socializioz reads the active brand, audience, connected accounts, campaign requirements, and available media.
  </Step>

  <Step title="Build the campaign">
    The Agent creates the campaign, content plan, and platform-specific post ideas.
  </Step>

  <Step title="Prepare the assets and posts">
    Media is generated, imported, or selected; posts are created and validated for their destinations.
  </Step>

  <Step title="Review and approve">
    The team edits, approves, or rejects the prepared work.
  </Step>

  <Step title="Schedule and publish">
    Approved posts enter the publishing queue and their returned state is verified.
  </Step>

  <Step title="Learn">
    Available campaign and post performance informs the next content cycle.
  </Step>
</Steps>

## External AI clients

The native [Socializioz MCP](/socializioz-mcp) lets supported AI clients use the same authorized workspace model through 38 tools. External clients do not bypass Socializioz permissions, approvals, plan rules, connected-account capabilities, or publishing validation.

## Related

* [Socializioz Agent](/features/ai-chat)
* [Campaigns](/features/campaigns)
* [Publishing Hub](/features/publishing-hub)
* [Integrations](/features/integrations)
* [Socializioz MCP](/socializioz-mcp)
* [Security](/account/security)