---
title: "Adaptive optimization"
slug: "adaptive-optimization"
abstract: "The Adaptive Optimization flow uses analytics data to identify content gaps and generate AI-driven recommendations for continuous improvement."
---
# Adaptive optimization

**Adaptive Optimization** is a flow that improves existing content using real analytics data and AI-driven recommendations. It identifies performance gaps, surfaces optimization opportunities, and suggests updates to keep content effective and relevant.

## When to use

- Content is underperforming
- You want data-driven SEO improvements
- You aim to increase engagement or conversions

## Agents in this flow

The **Adaptive Optimization** flow includes these agents:

1. **Researcher** - reviews your prompt and analyzes it against established UX and web design principles based on industry standards.
2. **Content Generator** - uses the research findings to create a blog that includes in-depth analysis and actionable recommendations.
3. **Translator** - optimizes the content and translates it into the selected languages.

## Run the flow

To run the Adaptive Optimization flow:

In the **Agents** section, the agents included in the flow are added by default. You can add more agents, replace, or remove agents as needed.

> [!important]
> Always review AI-generated artifacts carefully and validate key information before using them.

1. In the navigation menu, click **Agentic** and then, in the left sidebar, click :mdiPlaylistCheck: **Spaces**.
2. On the **Spaces** page, click **Create Space**.
3. On the **New space** page, in the **Space name** field, enter a name for the space and describe objective of your space.
4. In the **Starting point** section, click **Existing template** to use a predefined flow and, from the **Existing template** drop-down list, click **Adaptive Optimization**.
5. Click **Create space**.
6. For each agent, in the right pane as required, provide the necessary inputs:
    - **Parameters** - select the parameters to define how the content should be generated:
        - **Target Languages** - for example, *Spanish*, *French*.

        > [!tip]
        > To change the parameters the agent considers, you can edit the agent’s configuration.

7. Optionally, provide the following:
    - **Items** - use [items](https://doc.sitecore.com/sai/en/users/sitecoreai/working-with-agentic-studio/understanding-agents/run-an-agent/use-items-when-running-agents.html "Use items when running agents") as the inputs the agent will act on. The agent processes each item to generate outputs. Click **Add items** to include:
        - **CMS** - select content item from a SitecoreAI site or collection.
        - **Artifacts** - use artifacts previously generated in any space.
        - **Briefs** - select a [brief](https://doc.sitecore.com/sai/en/users/sitecoreai/plan-campaign-strategy/managing-briefs/managing-briefs.html "Managing briefs") created in the Strategy app.
        - **Other** - import a CSV file and select all or part of its data.
    - **Context** - use context as background information that guides how the agent generates results. Click **Add context** to include:
        - **Artifacts** - use any artifacts previously generated in the current space.
        - **Brand kits** - select your organization's [brand kit](https://doc.sitecore.com/sai/en/users/sitecoreai/design-components/brand-kits/brand-kits.html "Brand kits").
    - **Supporting files** - click **Browse** to upload supported files related to your request.
8. In the right pane, for the **Research Phase**, do the following:
    - If setup is required, click the **Research Phase** agent and enter a prompt to describe the topic you want to research (for example, *Social media marketing hacks for the Gen Z audience*).
    - Click :mdiPlayOutline: **Run** to start the agent.

    Processing might take some time. You can leave the page and return to the [space](https://doc.sitecore.com/sai/en/users/sitecoreai/working-with-agentic-studio/understanding-spaces/understanding-spaces.html "Understanding spaces") at any time to track progress. When processing is complete, the results appear in the space.

9. In the right pane, for **Content Generation**, do the following:
    - If setup is required, click the **Content Generation** agent and enter a prompt to describe your content requirements. For example, *Write a comprehensive blog post using the research* and click **Save**.
    - Make sure the correct channels, target audiences, and content formats are selected.
    - Optionally, add context or supporting files.
    - Click :mdiPlayOutline: **Run** to start the agent.

        The agent analyzes your prompt and generates an initial draft for review.

    - Edit or accept the draft to finalize the artifact generation.
10. In the right pane, for **Optimization**, do the following:
    - If setup is required, click the **Optimization** agent and provide the required input.
    - Enter or paste the content to translate (for example, *Develop account-specific case studies and one-pagers*.)
    - Make sure the correct languages are selected.
    - Optionally, add context or supporting files.
    - Click :mdiPlayOutline: **Run** to start the third agent.

        The agent analyzes your prompt and generates the requested content.

    - The number of artifacts corresponds to your selected languages.

To learn more about what’s included in a space, see [Open a space](https://doc.sitecore.com/sai/en/users/sitecoreai/working-with-agentic-studio/understanding-spaces/open-a-space.html "Open a space").