Instead of navigating the DynamicWeb UI to configure, set up, or manage your solution, you can instruct an AI agent in plain language — and it will do it for you. This is a fundamental shift in how DynamicWeb can be used.
Rather than clicking through settings, filling out forms, and manually executing workflows, you describe what you want. The agent understands your intent, determines the right steps, and executes them directly in DynamicWeb. This makes tasks that previously required deep platform knowledge and significant time-investment accessible to anyone who can describe what they need.
A simple example: creating a new product in DynamicWeb previously meant
- Navigating to the right shop
- Selecting the correct product group
- Filling out fields across multiple tabs
- Saving
Now you just tell the agent: "Create a product called Road Bike" — and it's done.
This also makes a new level of automation possible. Workflows that were too complex or time-consuming to automate before — like structuring a full PIM solution, enriching thousands of products, or setting up intricate discount rules — can now be driven by an agent operating continuously without manual intervention.
The DynamicWeb MCP Server
This is made possible by the DynamicWeb MCP Server, which acts as a universal connector so that any compatible AI-agent can interact with your DynamicWeb solution in a reliable and repeatable way.
The DynamicWeb MCP Server exposes platform functionality as a set of tools that AI agents can call. This means you can instruct an AI agent in plain language, and it will use the available tools to act on DynamicWeb on your behalf.
Note
MCP Server access is released in beta - future license for use maybe applied.
Connect an agent
Connect your agent using the MCP connection guides. The section includes shared prerequisites and setup guides for Claude, Codex, and GitHub Copilot.
Skills
Skills are pre-built prompt instructions that improve the precision and reliability of an AI agent working with DynamicWeb. Learn more in MCP skills.
Example use cases
Here are three use cases for agentic AU used with the DynamicWeb MCP server.
PIM implementation
Setting up a PIM solution involves identifying customer needs, designing a data structure, and creating it in DynamicWeb. Using the PIM Solution Skill, an AI agent handles the entire flow:
- The agent asks discovery questions about product types, shared attributes, type-specific attributes, and quality requirements
- It proposes a data model structure and waits for approval
- On approval, it builds the full structure, including data models, reference fields, category fields, completion rules, and workflows
This reduces the time from requirement to working PIM structure significantly.
Product enrichment
Once the PIM is running, day-to-day enrichment can be automated. Give an agent access to a specific product query and describe what needs enriching, such as SEO texts, product descriptions, or missing attributes. The agent reads the matched products, generates the content, and writes it back.
The agent can also assign products to the correct data models, ensuring that products are categorized correctly before or during enrichment.
Typical use cases include SEO title and description generation, long-form product copy, translation, completeness gap-filling, and data model assignment.
Commerce discount setup
Setting up complex discount rules manually is time-consuming. Describe your requirements to the agent in plain language, including which products, which customers, when the discount applies, and how it is calculated, and the agent will translate that into the correct discount rule configuration.
This is especially useful for multi-condition rules that would otherwise require many clicks and careful attention to rule ordering.
Note
Discount setup requires the New Discount Experience to be enabled in DynamicWeb