On 28 October 2026 Adobe published its agentic commerce update for Adobe Commerce. It covers three areas: an AI agent that helps developers build integrations, AI-assisted storefront authoring, and tools that help products show up correctly in answers from large language models (LLMs) such as ChatGPT. Most of it targets Adobe's paid cloud products, but the direction matters for every Magento 2 store, including Open Source shops on Hyvä. Here is what was announced and what you can do with it today.
1. Commerce Developer Agent is generally available
The Commerce Developer Agent takes a requirement written in plain language, turns it into a reviewable plan that Adobe calls a blueprint, and then scaffolds, validates and deploys the code as an Adobe App Builder app. In Adobe's launch blog post the example is a cart limit: the developer describes the rule, the agent builds a webhook, and the limit and the shopper message become settings a merchant can change in the Admin without another deployment.
Two details are worth knowing before anyone sells this to you as a robot developer:
- It works with App Builder, the out-of-process extension model. It does not write classic in-process Magento PHP modules for you.
- Adobe's own capabilities page says all generated code must be reviewed before production, calls the agent a starting-point generator rather than an autonomous developer, and asks for extra scrutiny on payment and security-sensitive logic.
- The same agent powers a Migration Assessment that scans an Adobe Commerce on Cloud (PaaS) codebase to plan a move to Adobe Commerce as a Cloud Service.
For agencies this mainly speeds up integration work and migration estimates. For merchants on Magento Open Source it changes little directly, because App Builder is part of Adobe's commercial offering.
2. Experience Workspace: storefront pages from prompts
Experience Workspace is in early access, with general availability planned for November according to the announcement. It lets teams build and edit storefront pages from natural-language prompts, with visual editing and live preview in one place, and it can be extended with skills, saved prompts, custom agents and MCP connections (MCP, the Model Context Protocol, is a standard way to connect AI tools to outside data). It is for storefronts on Adobe Commerce as a Cloud Service and Adobe Commerce Optimizer, so a Luma or Hyvä theme on your own server won't get it.
3. Product discovery on LLM surfaces is expanding
This is the part that affects every merchant, whatever platform they run. Adobe is adding several features by the end of November:
- PDP optimization: product page improvements aimed at helping LLMs match shopper intent to the right product, without changing what human visitors see.
- Catalog Enrichment: turning technical specs into richer product names and descriptions so AI can match products to queries.
- Brand Governance and Enrichment Manager: rules for tone and phrasing per SKU, plus scheduled, reviewed enrichment jobs for chosen categories.
Adobe's own architecture guide for agentic commerce frames the goal as one source of truth that serves both human shoppers and AI agents. The message is clear: AI assistants, whether ChatGPT or a chat window on your own store, can only recommend what your catalog data actually describes.
What a Magento 2 merchant can do now, without Adobe's cloud
You don't need an Adobe Commerce licence to act on the same idea. Most of the work is catalog hygiene you control in the Magento Admin:
- Fill in the attributes shoppers actually ask about (size, material, compatibility, use case) as real attributes, not only inside description HTML.
- Rewrite product names that are just internal codes so they say what the product is.
- Make sure configurable products have clear option labels and that price and stock are current, because an assistant that recommends an out-of-stock variant loses trust fast.
- Review store views separately if you sell in several languages.
Our earlier checklist of Magento 2 attributes for AI product discovery goes through this in more detail.
Where Claspwell fits
Claspwell is a catalog-aware AI sales assistant for Magento 2 stores on Hyvä, Luma and Blank themes. It syncs your catalog and answers shoppers from your real products, prices and stock, rather than from a generic model's guesses. It doesn't depend on App Builder or Adobe's cloud services, and a typical install takes 3 days to 1 week. The same catalog work described above that helps you show up in LLM answers also makes an on-site assistant more useful.
Source: Adobe Commerce unlocks agentic commerce, from code to checkout, Adobe partner news, 28 October 2026.