Intelligent Selling Services in SAP Commerce Cloud: AI-Powered Personalisation

Customers don’t think about “personalisation”, and they don’t care about your recommendation engine. They leave when the products on screen feel random. Intelligent Selling Services (ISS) is SAP’s answer to that for Commerce Cloud: a cloud service that captures shopper behaviour in real time and uses it to personalise product recommendations and merchandising.

Below: what ISS does, how its modules work, how it relates to Adaptive Search, and what it needs before it delivers results.

TL;DR: Intelligent Selling Services (formerly Context-Driven Services) is a cloud service alongside SAP Commerce Cloud. It captures clickstream and purchase behaviour and uses it for one-to-one product recommendations and merchandising carousels, with reporting and A/B testing to prove the effect. Search ranking is handled by Commerce Cloud’s Adaptive Search module, which merchandisers configure with boosts and rules. ISS needs traffic and clean product data to learn; low-traffic shops see smaller gains.

What is Intelligent Selling Services?

ISS is not a recommendation widget you bolt on. It is a separate SAP cloud service that runs outside the Commerce Cloud application server, formerly known as Context-Driven Services (CDS) (SAP Help: Intelligent Selling Services). The storefront sends behavioural events (clicks, searches, cart additions, purchases) to ISS, ISS processes them, and returns personalised results via API. Because the heavy lifting happens outside the shop, it does not slow down your storefront.

ISS is organised in modules:

1. Foundation: captures and stores behavioural data and customer profiles
2. Recommendation: one-to-one, behaviour-driven product recommendations for known and anonymous customers
3. Merchandising: product carousels driven by strategies that combine business goals with behaviour
4. Reporting: KPIs for the carousels, including the results of A/B tests

Whether ISS is part of your Commerce Cloud subscription depends on your contract, so check your order form before planning a project around it.

How do product recommendations work?

ISS offers several recommendation approaches, each suited to a different situation:

Collaborative filtering. The classic “customers who bought this also bought…” approach, based on purchase patterns across your customer base. The more transactions you have, the better the recommendations become.

Content-based filtering. Recommends products with similar attributes (category, brand, price range) to what the customer viewed or bought. It works even with little purchase history: a first-time visitor who has looked at three products already gets reasonable suggestions.

Context-aware recommendations. What you recommend depends on where the customer is. A recommendation on the cart page (cross-sell) is a very different thing from one on the home page (discovery).

Trending products. Surfaces products with rising demand: seasonal items, new launches, products that suddenly take off.

The models update continuously as new data arrives; nobody has to retrain them manually.

How does ISS relate to Adaptive Search?

This is often mixed up. Search ranking in Commerce Cloud is handled by the Adaptive Search module, part of the Commerce platform rather than ISS (SAP Help: Adaptive Search Module). Merchandisers use it to define search profiles per catalogue, category or customer segment, with facets, boost rules, boosted and buried items and sort orders.

Together the two cover the main levers:

Search and category pages. Adaptive Search lets you boost new arrivals, promote high-margin products or push out-of-stock items down, globally or for specific categories and segments. Category pages use the same search configuration, which is where more revenue hides than most commerce teams expect.

Carousels and recommendations. ISS fills the product carousels on home, product, category and cart pages with personalised or strategy-driven products.

Human judgement plus machine learning. Merchandisers set the business rules; the behavioural models do the per-customer ranking within them.

What does merchandising with ISS do?

In practice, merchandising is where ISS earns most of its keep:

Strategy-driven carousels. Merchandisers define strategies (for example “similar products”, “frequently bought together”, “trending in this category”) and ISS selects the products per customer and context. Manual pinning remains possible.

Business rules. Boost brands, exclude products, respect margins or stock levels. The AI works inside the rules you set, not around them.

A/B testing. Compare strategies, or AI-driven carousels against manual curation, and measure which one converts better for your catalogue and customers. That is how you find out whether personalisation pays off in your shop, rather than assuming it does.

What results can you expect?

The effect of ISS depends on catalogue size, traffic, data quality and how mature your personalisation was before. We therefore do not quote generic uplift figures. What we do in projects is define a baseline (for example click-through on a static “top sellers” carousel, cart value, search-to-cart rate), switch on one ISS strategy, and compare with an A/B test.

Two conditions decide the outcome. First, traffic: the models learn from behaviour, so low-traffic shops see smaller gains. Second, product data: recommendations based on attributes are only as good as the attributes. If you are thinking about the wider AI picture beyond commerce, our overview of what is real and what is roadmap in SAP CX AI helps sort the options, and the storefront side of AI is covered in our post on the SAP Commerce Cloud MCP server and agentic commerce.

To discuss whether ISS fits your shop, talk to our Commerce team or visit our SAP Commerce Cloud page.

FAQ

Is ISS included with Commerce Cloud?

ISS is a separate SAP cloud service that works alongside SAP Commerce Cloud. Whether it is part of your subscription depends on your contract, so check your order form or ask SAP before you plan the rollout.

How much traffic does ISS need to be effective?

There is no fixed threshold, but ISS learns from behaviour, so the more sessions and orders flow through the shop, the better the recommendations become. Shops with little traffic benefit more from content-based strategies and merchandiser rules than from collaborative filtering.

Can I control what ISS recommends?

Yes. Merchandisers define the strategies and business rules: boost brands, exclude products, pin items to specific slots or respect margin and stock limits. The AI personalises within those rules.

Does ISS work with Composable Storefront?

Yes. Composable Storefront has an integration library for ISS that provides the tracking and the merchandising carousels as standard components (Composable Storefront documentation). More on the storefront itself in our Composable Storefront guide.

How does ISS compare to third-party recommendation engines?

ISS’s advantage is its close integration with Commerce Cloud: the same catalogue, customer data and order history, with no extra data synchronisation. Third-party engines such as Algolia, Nosto or Dynamic Yield may offer more algorithms or cross-platform use, but they need additional integration and data sync. In short, ISS wins on simplicity, third-party engines on flexibility.