Magento / Adobe Commerce Customer Support Guide

Why Magento support is complex

Magento (Adobe Commerce) is chosen by merchants who need things that hosted platforms cannot provide: deep customization, complex product types (configurable, bundle, grouped, downloadable, virtual), multi-website setups with different catalogs and price rules, and extensive B2B functionality including company accounts, shared catalogs, and purchase orders.

Each of these features creates a corresponding support complexity. Configurable products generate more 'I ordered the wrong variant' tickets. Multi-website setups mean different policies per site. B2B purchase orders create invoicing questions that smaller B2C stores never see.

The answer is not to avoid automation — it is to build an agent that is connected to your specific Magento data and configured to handle your specific ticket mix.

What to automate on Magento

Despite the added complexity, the core high-volume ticket categories on Magento stores are the same as everywhere else. Start there before tackling the platform-specific edge cases.

API integration options for Magento

For headless Magento setups (PWA Studio, Hyvä, custom frontends), the widget is added to the frontend framework rather than the Magento theme. See our headless guide for specifics.

1. 1In Magento Admin, go to System > Extensions > Integrations and create a new integration.
2. 2Assign resource permissions: Sales > Orders (read), Customers (read), Catalog > Products (read). Add write permissions if the agent will process RMAs or refunds.
3. 3Activate the integration and copy the Access Token.
4. 4In Bookbag, choose Magento/Adobe Commerce and paste the store URL and Access Token.
5. 5Import your return policy, FAQ, and product guides as knowledge documents.
6. 6Add the Bookbag widget to your Magento theme via a layout XML block or a custom module.

Multi-website and multi-store handling

Magento's multi-site architecture allows a single installation to run multiple websites, each with multiple stores and store views. This maps to different catalogs, price rules, shipping options, and return policies per website.

When configuring AI support for a multi-site Magento setup, use a separate agent configuration per website (or at least per policy scope). Route chat traffic based on the URL the customer is on, and pass the store code with each API request so the agent retrieves the right data for that store.

• Create separate Bookbag workspaces for each Magento website that has distinct policies.
• Pass the store code or website ID in API requests to scope order and product lookups correctly.
• Use different chat widget snippets per website so the correct agent loads.
• Consolidate reporting across sites for a global view, but drill down per-site to optimize each.

B2B features in Adobe Commerce

Adobe Commerce's B2B module adds company accounts, shared catalogs, quick order, requisition lists, and negotiable quotes. These features generate support questions that do not exist in B2C stores.

For B2B support automation, the most valuable automations are: order status for purchase orders, invoice copies (the agent can retrieve and email the order invoice PDF), and shared catalog product availability checks. Negotiable quote questions require human involvement unless the quote stage is very early.

Knowledge base strategy for Magento stores

Magento does not have a built-in knowledge base or FAQ system. Most merchants use a CMS page structure or a third-party help center. Export your policy pages and FAQ content as plain text for the agent's knowledge base.

For complex product catalogs — especially configurable products with many attributes — consider creating a supplementary product guide document that maps common customer questions ('Does the XL fit a 36-inch waist?') to answers. The agent cannot infer this from SKU data alone.

• Export all CMS pages tagged as policies or FAQs and import to Bookbag.
• Create a 'product types explained' document if your catalog includes bundle or grouped products that confuse buyers.
• Add a guide for your RMA process — the steps to initiate a return, what documentation is needed, and turnaround times.
• For multi-website setups, maintain separate knowledge documents per site with site-specific policies.

Deployment and testing on Magento

Magento stores typically go through a more rigorous change management process than smaller platforms. Before going live with AI support, run a structured test plan that covers your most common ticket types with real order data.

Test on staging first. Create test orders in your staging environment and verify the agent can look them up, apply the correct policy, and escalate appropriately. Check that the widget loads correctly on all device types and that multi-site routing works as expected.