Shifting the Search Paradigm
Last year, we introduced Content AI as a pivotal step toward redefining the search paradigm—with our mission to vectorise all AEM managed content and bringing semantic search capabilities alongside traditional lexical approaches. This marked the beginning of a more intuitive and context-aware search experience.
Over the past year, we have significantly expanded the scope and adoption of this service. In this session, we will showcase how Content AI is now powering a diverse set of solutions in and around the AEM ecosystem. We will highlight key features, evolving capabilities, and real-world integrations—from Content Hub use cases that demand a combination of semantic and advanced lexical search, to Brand Concierge experiences enabling natural language-driven discovery.
Additionally, we will share progress on upcoming enhancements, Natural Language Search, support for Oak indexes within Content AI APIs, enabling more efficient and scalable search across structured content.
Join us to explore how we are continuing to transform search into a more intelligent, unified, and developer-friendly experience across Adobe.
alexandereggers
As of now, do you still need to manually set up what specific bits of information to look for when evaluating a page for a vector for the LLM search, or can this too be automated?
bnecula
If you mean deciding which information from a page should be used for embeddings/vector search, that’s already part of the flow, it all happens behind the scenes as part of how vector search works. You don’t need to manually specify which bits of information to look for on each page; the content processing, embeddings, and retrieval are handled automatically. If you mean evaluation datasets for measuring search quality, that’s a different process altogether. Happy to share more details on how that works if that’s what you’re referring to.
Cornelius
Does it work with EDS UE?
MOHIT KATARIA
We are working on EDS support. It will be available by next month.
Konrad Windszus
Is AEM Content AI now GA? How is it licensed?
Konrad Windszus
API (https://developer.adobe.com/experience-cloud/experience-manager-apis/api/experimental/contentai/) still states: > Note : The use of these APIs is subject to co-innovation arrangements with Adobe or licensing once available for public use. Please contact your Adobe representative for more details.
Constantin Ioan-Mihai
How does this solution differ from Brand Concierge? Is this for authoring purposes only?
MOHIT KATARIA
Content AI is retrieval and augmentation on top of aem content, a foundational subset of BC. Making use of CAI APIS can serve any usecase that requires retrieval, content intelligence, augmentation/generative.
Adrian
Does this also work with DA?
MOHIT KATARIA
DA sites will be supported by end of october. For now we can support it with manual intervention. Feel free to connect, we can work in co-innovation model.
Peter P
How does indexing looks like and how can we browser index if we look at traditional lucene based indexes? Can we see the index?
MOHIT KATARIA
For content-ai we can list source config. These source configs do have details on how content is indexed. You can have a look at content-ai apis. Refer: https://developer.adobe.com/experience-cloud/experience-manager-apis/api/experimental/contentai/