AI Impact Stack · Composable Commerce
Your second market costs less than the first.
A composable platform connects shop, ERP, search and content via APIs. New markets, brands and channels are added by configuration, and agents find assortment, prices and stock where they need them.
Composable commerce in the AI Impact Stack
Every building block selectable, replaceable and reachable for agents.
An agent that advises, reorders or creates orders needs the same capabilities as your shop: reading the assortment, checking prices and availability, triggering cart and order. In a composable platform, these capabilities are available as independent services via APIs. In many grown shop systems, they sit in templates and extensions that only the shop itself can use.
That way, every new use case, whether purchase advice, reordering in the B2B portal or visibility in AI assistants, builds on the existing platform. New markets, brands and channels are added by configuration. A change in one place stays in that place, and releases become smaller and more frequent.
We build new platforms on commercetools, for B2C, B2B, D2C and app. We have also built platforms with Spryker and SCAYLE, for example for Segmüller and the Witt Group. We connect ERP and PIM so that they remain the leading systems.
What the platform delivers
The difference rarely shows on launch day. It shows with the second market, the third channel and the first requirement that nobody knew about during the build.
Brands, countries and tenants run on a shared codebase. Witt Group runs 20 shops on one platform, and Möbel Boss uses the multi-brand platform of porta Möbel.
Result: every additional market launches faster and with less effort than the first.
Campaigns and content go live as soon as they are ready, independent of the release date. You choose specialized services for search or loyalty per building block.
Result: shorter paths from idea to revenue.
Assortment, prices and stock are available via APIs, for buying advice, reordering in the B2B portal and visibility in AI assistants.
Result: AI use cases go into production on the existing platform.
Six to nine months.
That is how long a composable rebuild typically takes, from requirements to ERP and PIM integration, at the lower end with a tightly cut scope. The pet supplies retailer was live after six months.
Technology partners
We build new commerce cores on commercetools, complemented as needed with content, search and delivery, running in the cloud you bring.
References
We have worked with porta Möbel since 2019, from the legacy analysis to the composable architecture. Möbel Boss uses the same multi-brand platform, and Witt Group runs 20 shops on one platform.
Since 2019, foobar Agency has accompanied porta Möbel on its path to digital market leadership: from legacy analysis via incremental headless commerce to a full composable commerce architecture on MACH principles.
As a sister brand of porta Möbel, SB Möbel Boss benefited from the jointly developed multi-brand platform. The replatforming from SAP Commerce to Commercetools was completed – with a tailored storefront and on time before Black Friday 2024.
20 shops on high-maintenance in-house systems – too expensive, too slow. foobar Agency developed a cross-brand design system and replatformed all shops to SCAYLE and Next.js.
FAQ
For companies with real complexity. If three or more of these points apply, the approach pays off: several markets, brands or tenants; pricing logic in the ERP; more than one frontend, such as shop, partner portal and app; business teams that should work independently of releases; specialized requirements for search or loyalty; AI initiatives with a production goal.
For a single market with a stable assortment, an integrated system remains more economical. For a specific symptom such as weak search, a targeted measure in the existing system helps. If product data is still being built up, the data foundation comes first and the platform after.
We build on what works. Existing systems, services and patterns are reused and connected. We renew where a system blocks the value lever.
In four phases with defined results. We offer the architecture review (2 to 4 weeks) and discovery (3 to 5 weeks) at a fixed price. They are followed by MVP to release candidate in three to four months, and a rollout via a pilot market with subsequent operations.
By making assortment, prices and availability machine-readable via APIs. Your own agents for advice and reordering build on this, and your customers' AI assistants find your offering. Via MCP servers, agents also reach the ERP and internal services.
The Impact Check assesses your systems along the four hurdles and shows which platform your first agent will build on.
We look forward to your enquiry.
Please accept marketing cookies to load the registration form.