> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ones1ght.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Analytics / Actions

> Analyze store data in natural language and turn the answers into actionable policies.

## Analytics / Actions

**Analytics / Actions** is a conversational screen where you ask questions about your store data in natural language, and the AI returns root-cause analysis together with recommended, actionable policies.
A **"Recommendation L2 · Execute after approval"** badge appears at the top: the AI only proposes policies, and actual execution happens after a member approves.

<Info>
  This page is visible to every member of your organization (Owner, Manager, and Member).
</Info>

## Analysis flow

<Steps>
  <Step title="Select a Skill">
    Choose the OneS1ght Skill used for analysis. The default is **"Sales Equation Analysis"**,
    and the **＋ Store** button opens the [Skill Store](/en/agent/skill-store) to install a new Skill.
  </Step>

  <Step title="Select an AI platform (BYOK)">
    Choose the AI platform that runs the analysis — Claude, GPT, or Gemini. Each shows a key-status
    badge (**Key connected / Server key / No key**). Platforms with no key are disabled; manage keys
    in [Settings](/en/agent/llm-models).
  </Step>

  <Step title="Enter your question">
    Type your question in natural language. You can attach **up to 5 files**; text files have their
    contents included in the analysis, while other files contribute their filename only.
  </Step>

  <Step title="Register a policy">
    Register a policy from the answer (see below).
  </Step>
</Steps>

<Note>
  Responses received without a connected BYOK key are marked with a **"Sample response"** badge.
</Note>

## From answer to policy

You can register two types of policy directly from an AI answer.

| Type                        | Target                        | Action                                                                     |
| --------------------------- | ----------------------------- | -------------------------------------------------------------------------- |
| 🏢 B2B store-manager policy | Store operations improvement  | Register with an Action Agent → confirm the rule                           |
| 📱 B2C customer-app policy  | Persona-based personalization | Delivered via the customer app (SDK) → register as "Provide to smartphone" |

Example personas: teens / women 20s–30s / returning / new.

## Recommendation → Execution (right panel)

Handle recommended policies in the **"Recommendation → Execution"** panel on the right.

<CardGroup cols={2}>
  <Card title="Approve and execute" icon="circle-check">
    Approving a recommendation moves the policy into execution.
  </Card>

  <Card title="Edit" icon="pen">
    Edit the recommendation before registering it.
  </Card>
</CardGroup>

Policy status: **proposed / approved / dismissed**.