AI Agent

Accelerate time-to-impact with AI Agent

AI Agent helps you get real work done in Kontent.ai. Unlike AI chatbots that stop at suggestions, AI Agent breaks down your tasks into steps, navigates your content model, executes changes, and shows you the results. Whether you need to audit existing content, perform bulk operations, or create new content at scale, AI Agent handles the complex tasks while you focus on strategy and outcomes. With AI Agent by your side, here’s a non-exhaustive list of what you can do:

• Find outdated content items and create a plan for improving SEO and GEO
• Review and edit a content item and apply the same changes to all existing language variants
• Draft content items based on your inputs
• Translate and publish items in different necessary languages

Tune in to the video below to hear what the agent can do.

Get access to AI Agent

Experience the power of the AI Agent firsthand by enabling it in Innovation Lab. If you don’t see the agent yet, reach out to Support to get access. Once enabled, you’ll see the AI Agent on the left-hand side of the navigation and witness the transformation of content management.

How to use AI Agent

As you chat with the agent, here are some example prompts you can customize to match your needs.

Audit and optimize content

Perform content audits and optimize your existing content:

• “Find all articles about coffee published before January 2026 that don’t mention our new sustainable sourcing initiative. Show their current workflow status.”
• “Analyze content gaps in our healthcare industry coverage by comparing our existing articles against the top 10 healthcare topics from our competitors’ websites: {List of competitors}. Suggest 5 specific new topics with target keywords and potential content angles.”
• “Update all product descriptions in the Electronics collection to include our new brand voice guidelines: use active voice and highlight sustainability features in the first sentence. Show me a preview of the first 3 changes before applying them to all items.”
• “Find duplicate images in my asset library by comparing file content (not just filenames). Group duplicates by resolution and file size, then suggest which versions to keep.”

Explore what’s in your content

Ask what’s in your content model to understand what it can access:

• “List all content types in my project grouped by their usage count. Show me which types have content items and which are empty.”
• “Show me the elements in my Article content type, whether they’re required and have limitations.”
• “Find all content type snippets I have, list which content types use each snippet, and identify any unused snippets.”

Find and replace what‘s needed

Find what you need and make changes in your content to align with your brand guidelines, rebranding, and so on:

• “In all articles with the Documentation taxonomy, replace ‘users’ with ‘developers’ but keep ‘end-users’ unchanged. Show me the first 5 replacements for review before applying to all items.”
• “Find all mentions of our old product name ‘A’ across content items, assets, and URL slugs. Update them to our new product name ‘B’ while preserving URL redirects. Create a report showing all changed items.”
• “Update all content using the taxonomy term Customer Support to replace the text ‘contact support at [email]’ with ‘start a chat with our support team using the chat widget in the bottom right corner’. Apply this change only to Published items.”

Model and create content in seconds

Transform ideas into structured content:

• “Create a blog post about sustainable packaging trends for 2025. Include 3 specific trends with data points, 2 case studies from major brands, and a conclusion with actionable recommendations. Use the Blog post content type and tag it with Sustainability and Industry Trends taxonomies.”
• “Design a content type called Product Review with these elements: Rating (number 1-5), Pros (rich text, required), Cons (rich text, required), Summary (text, 160 characters max, required), Reviewer Name (text, required), Review Date (date & time, required), and Product Category (multiple choice taxonomy).”
• “Based on this document, identify which existing content types match the structure. If the document contains information that doesn’t fit our current content types, suggest new content types with specific elements. Then create draft content items for each product mentioned in the document.”

AI Agent understands your content model and ensures new content follows your requirements.

Perform bulk operations

Do mass content changes through requests:

• “Schedule all articles by John Roy that are in the Review workflow step to publish on September 30, 2025 at 9:00 a.m. UTC.”
• “Update all content items in the Product Description content type to include sustainability information in the new Sustainability Features element. Research and add 2-3 relevant sustainability attributes based on the product category.”
• “Tag all articles that mention ‘machine learning’, ‘neural networks’, or ‘LLM’ with the Artificial Intelligence taxonomy term. Exclude articles already tagged with Software Development. Generate a report showing how many articles were newly tagged.”

Research to support your content

Search the web for current information to support your current strategy:

• “Research the latest trends in sustainable fashion from the past 3 months. Create a blog post that explains each trend with specific brand examples and end with how our products align with these trends.”
• “Find recent developments in renewable energy during Q4 2025. Compare these developments against our existing renewable energy articles and suggest which content needs updating.”
• “Look up space events happening in October 2025. For each event, create a News Article content item that includes event details and its scientific significance. Set each article to Review workflow step.”

Review AI Agent changes

After the agent performs bulk updates to your content items, you can verify the intended changes directly in the chat. You’ll see an expandable bar showing the content item’s name. Expand to view only the elements that were modified with highlighted changes. These changes are also found in the item’s version history. This focused view helps you quickly assess the agent’s work before publishing.

Considerations

When working with the AI Agent, keep these points in mind:

• AI can make mistakes. Always review the changes made to your content.
• AI Agent uses your subscription’s allocated API calls and bandwidth to perform operations.
  • Most requests typically consume multiple API calls and bandwidth, while bulk operations require resources proportional to the number of items processed.
• Only your latest conversation in each project is preserved locally in your browser.
• While the agent understands conversational requests, being specific and using Kontent.ai terminology in your request yields better results.
• Changes made by the AI Agent are saved in version history, but they currently appear under the name of the contributor who started the action. In the future, we plan to make it clearer when changes are made by the agent.
• The AI Agent works within your existing content model and can only make changes according to your instructions.
• AI Agent respects your established workflow steps, roles, and permissions, accessing only what you‘re allowed in the project.
• The AI Agent cannot activate or deactivate languages. New languages are always created as active. To change a language’s status, use the Kontent.ai UI.
• When working with multilingual content, specify the target languages clearly.
•  Large-scale changes are processed systematically. AI Agent will inform you of its progress and any issues encountered during bulk operations.
• The AI Agent relies on the Management API, which does not support comments and suggestions, so any comments in elements that the agent edits are resolved. This is a known limitation to consider when comments are used in content items that may be modified by the agent.
• AI Agent doesn’t use your data for training. Your content remains secure and private.

Help shape what comes next

Your input helps us improve this capability. If you have any suggestions, questions, or feedback, we'd love to hear from you. Share your experience using the button in Kontent.ai or in the Learn portal.