Adobe Commerce

Adobe Commerce's real AI features run through Adobe Sensei (Adobe's shared AI/ML layer across its products) powering product recommendations and Live Search, which ranks and personalizes search results based on behavioral signals rather than plain keyword matching -- this is genuine, product-specific AI, not a generic marketing claim layered on top.

Practical Details

These features need real usage data to perform well -- a newly launched catalog with little purchase or browsing history will see weaker recommendation quality at first, since the models improve as they accumulate signal, not on day one.

Actionable Considerations

1. Expect recommendation and Live Search quality to improve over time as real customer behavior data accumulates, not to be strong immediately on a new catalog.
2. Use Live Search specifically for its behavioral ranking, distinct from the platform's older, more literal catalog search.
3. Review and validate AI-generated recommendation placements periodically, rather than assuming the model's choices are always merchandising-appropriate.

Full Page Cache (built on Varnish) is essential infrastructure at real production scale, not an optional optimization -- without it, catalog and content pages regenerate on every request, which becomes a genuine bottleneck under real traffic. Multi-Source Inventory (MSI) handles stock across multiple warehouses or fulfillment sources with priority-based allocation, which matters once a merchant outgrows a single stock location.

Practical Details

Adobe Commerce Cloud's separate integration, staging, and production environments exist specifically so scaling-related infrastructure or configuration changes can be tested before they hit live traffic, rather than validating changes in production.

Actionable Considerations

1. Confirm Full Page Cache (Varnish) is properly configured before assuming performance issues are a code or hosting problem elsewhere.
2. Set up Multi-Source Inventory once you're fulfilling from more than one location, rather than tracking stock as one undifferentiated pool.
3. Test infrastructure and configuration changes in a staging environment before applying them to production, especially on Adobe Commerce Cloud's environment structure.

Customer and order data segregation on a multi-website Adobe Commerce install follows the Website/Store/Store View structure -- data can be scoped so customers and orders from one website aren't visible in another, which matters for merchants running genuinely separate brands or regions from a single installation.

Practical Details

Privacy compliance (like GDPR data subject requests) isn't a single built-in toggle -- it typically involves a combination of platform configuration, custom development, or a dedicated privacy extension, so treat specific compliance requirements as a real implementation task rather than assuming out-of-the-box coverage.

Actionable Considerations

1. Plan website/store scoping deliberately if data segregation between brands or regions is a real requirement, since it's structural, not a simple setting.
2. Don't assume a specific privacy regulation (like GDPR) is fully handled out of the box -- confirm what configuration or extension work is actually needed.
3. Document data retention and access policies explicitly, since Adobe Commerce doesn't impose them for you by default.

A foundational early decision is deployment model: Adobe Commerce Cloud (Adobe's managed hosting, with a git-based workflow across separate integration, staging, and production environments) versus an on-premise license you host and manage yourself -- these are genuinely different operational models, not just a hosting preference.

Implementation Details

Catalog and store structure follow a Website > Store > Store View hierarchy, which is how Adobe Commerce handles multiple brands, regions, or languages from one installation -- getting this structure right early matters, since restructuring it after significant catalog data exists is real, disruptive work.

Actionable Considerations

1. Decide between Adobe Commerce Cloud and on-premise hosting early, since it shapes your entire deployment and update workflow going forward.
2. Plan your Website/Store/Store View structure around actual brand, region, or language needs before loading real catalog data.
3. Use the Cloud environment's integration/staging separation to test changes before they reach production, rather than editing production directly.

For B2B merchants specifically, real engagement often runs through company accounts, negotiated quotes, and requisition lists -- workflows genuinely distinct from typical B2C browsing and checkout, and a real differentiator versus platforms built primarily for consumer storefronts.

Implementation Details

Realistic engagement measurement should track these B2B-specific flows (quote-to-order conversion, requisition list reuse) separately from standard B2C metrics, since lumping them together obscures what's actually driving revenue in a mixed B2B/B2C catalog.

Actionable Considerations

• Track B2B-specific metrics (quote conversion, requisition list usage) separately from standard B2C funnel metrics if you run a mixed catalog.
• Use company accounts and shared catalogs specifically for B2B customers who need multi-user purchasing with shared visibility, not as a general feature for all customers.
• Review negotiated quote turnaround time as an operational metric, since a slow quote process is a common real friction point in B2B sales cycles.

Adobe Commerce's GraphQL API is the modern integration layer, purpose-built for headless storefronts (like those built with PWA Studio) and generally preferred over the older REST and SOAP APIs for new development, though those legacy APIs remain available for existing integrations.

Practical Details

For deeper platform extensibility -- connecting to Adobe's broader Experience Cloud or building custom serverless extensions -- Adobe I/O and App Builder are the relevant tools, distinct from a standard third-party connector; most ERP and marketplace integrations in practice are built or configured by a Magento/Adobe Commerce System Integrator rather than installed as a simple plug-and-play app.

Actionable Considerations

1. Build new integrations against the GraphQL API rather than the legacy REST/SOAP APIs, unless you're extending an existing REST-based integration.
2. Budget for System Integrator involvement on non-trivial ERP or marketplace connections, since these are rarely simple one-click integrations on this platform.
3. Use Adobe I/O and App Builder specifically if you need custom serverless extensions or deeper Experience Cloud connectivity.

The real distinction worth understanding here is between Magento Open Source (free, community-supported only, no contractual SLA from Adobe) and Adobe Commerce (licensed, with contractual support and SLA-backed response times) -- teams sometimes assume they have Adobe support simply because they're running "Magento," which is only true on the licensed Adobe Commerce edition.

Practical Details

Security patches are released periodically and need to be applied deliberately -- an unpatched, outdated Magento/Adobe Commerce installation is a genuinely common real-world security risk, since it's a widely-targeted platform precisely because so many stores run it.

Actionable Considerations

1. Confirm whether you're actually running Magento Open Source or licensed Adobe Commerce, since only the latter includes Adobe's contractual support.
2. Apply security patches on a defined schedule rather than deferring them, given how frequently this platform is targeted precisely because of its popularity.
3. Test patches and updates in a staging environment before production, since Adobe Commerce updates can have real compatibility impacts on custom code and extensions.

Admin access uses a resource-based Access Control List (ACL) -- roles are defined by which specific admin resources (catalog management, order management, a specific report, etc.) they can access, letting you build genuinely narrow roles rather than an all-or-nothing admin account for every team member.

Practical Details

For multi-website setups, access and catalog visibility can also be scoped by website or store view, which is a separate layer from admin resource ACLs -- worth understanding as a distinct control, not the same setting, when a multi-brand or multi-region store needs both role-based and site-based access restrictions.

Actionable Considerations

1. Build admin roles around specific resource permissions your team members actually need, rather than defaulting to full admin access.
2. For multi-website installations, apply website/store-view scoping as a separate control layer from resource-based ACLs.
3. Review admin role assignments periodically, since accumulated broad access is a common, avoidable security gap on this platform.

Adobe Commerce (built on Magento's open-source core, which Adobe acquired in 2018) uses an EAV -- Entity-Attribute-Value -- data model for products, which is what makes wildly different product types (a t-shirt with size/color, a digital download, a configurable bundle) coexist in one flexible catalog without a rigid fixed-column schema. The real tradeoff is that EAV's flexibility comes at a performance cost from all the joins it requires, which is why catalog indexing and caching matter more here than on simpler platforms.

Key Benefits

Adobe Commerce (versus the free Magento Open Source edition) adds B2B features -- company accounts, shared catalogs, negotiated quotes, requisition lists -- plus Page Builder, a drag-and-drop tool for building merchandising content and landing pages directly in the admin without a developer.

Implementation Considerations

• Understand that Adobe Commerce (licensed, with B2B features and support) and Magento Open Source (free, community-supported) are genuinely different products, not just different names for the same thing.
• Plan your catalog attribute structure carefully upfront, since EAV makes changing core product attributes later a bigger operation than in a simpler fixed-schema system.
• Use Page Builder for merchandising content that marketing needs to update without developer involvement.