Diagram separating built-in platform capability from custom development work

Ask what AI Adobe Commerce gives you and you get two unhelpful answers. The vendor page says everything. A developer who has been burned says nothing.

The truth is a line, and it is worth knowing exactly where it sits, because on one side is a checkbox and on the other is a quarter of engineering work.

What actually arrives with the licence

Product Recommendations. A first-party module, powered by Adobe Sensei, that watches behaviour on your store and serves recommendation units. It is genuinely built in, you configure recommendation types rather than train anything, and it works without a data scientist. This is the one people mean when they say Adobe Commerce has AI.

AI-powered search. Adobe markets semantic search that reads intent rather than matching strings, plus automatic product ordering on category pages aimed at revenue rather than alphabet. Both are platform features.

That is close to the whole list. It is not nothing — recommendations and search are two of the three places personalisation actually shows up for a customer — but it is a shorter list than the marketing implies.

What belongs to a different product

Generative content creation, AI merchandising tooling, built-in A/B testing: these are Adobe Commerce Optimizer features. Optimizer is a separate product that Adobe positions as something you can run alongside an existing storefront without migrating.

This matters when you are budgeting. An article that says "Magento does GenAI content" and a quote that assumes those features are included will not survive contact with the licence. Ask specifically which SKU a capability sits in.

What you build

Three things get attributed to the platform that are not in it.

Chatbots. No conversational assistant ships with Adobe Commerce. Every store with one has either an embedded third-party widget or a custom integration against an LLM API. That is entirely doable, and it is a project with a scope and a bill, not a setting.

Fraud detection. Adobe Commerce integrates fraud services rather than shipping its own model. In practice that means selecting a vendor, wiring the integration, and deciding what the store does with a risk score at checkout — which is a business rule nobody else can write for you.

Dynamic pricing. Adobe Commerce has price rules, catalogue price staging and customer segments, and those cover a lot. Continuous algorithmic repricing against demand and competitors is not a stock feature. It is a service you build, feed and monitor.

Why the distinction is the whole point

A platform decision made on the assumption that AI is included, when half of it is a separate SKU and the other half is custom development, produces the same outcome every time: a roadmap that slips two quarters and a conversation about scope nobody enjoys.

The useful version of the question is not "does Adobe Commerce have AI". It is:

Which of these capabilities do we need in the first release, and which are year two?

For each one, is it in the licence we are buying, in Optimizer, or ours to build?

If it is ours to build, who maintains it after launch, and against what data?

The honest answer for most merchants is that Product Recommendations and better search cover the near-term win, and everything past that is a project. Which is fine. It is just better to know that before the budget is signed than after.

If you are working out which side of that line your requirements fall on, we are happy to go through it with you.