Why Nuxeo is more AI-ready than people think – For Dummies

By Derick Deleo

Can we be honest here? If you are not talking about AI-enabling Nuxeo within your organization you can be certain that someone is, and it is most likely someone who knows nothing about Nuxeo. AI product companies are bombarding organizations with marketing emails, sales emails, LinkedIn posts, enticing people to talk to them and then convincing those people that only they can AI-enable Nuxeo for your organization. But is that true? Do you really need to purchase some bolt-on solution to AI-enable Nuxeo?

My colleague and friend, Mariana Cedica, recently published a blog titled “Why Nuxeo is more AI-ready than people think” and we both thought that perhaps a non-technical version of that blog would help people who do not REALLY know Nuxeo understand that Nuxeo IS truly AI-ready, and it is easier than most people think to AI-enable Nuxeo. As Mariana describes in her blog, Nuxeo and the underlying components easily support storing AI generated embeddings. What are embeddings you ask? Well, think of embeddings as a mathematical representation of images, audio, words,...basically the content within your DAM or ECM system. These mathematical representations, or vectors, are designed to capture the semantic meaning and context of data, allowing machine learning models to understand relationships between items. For example, an image of an English Bulldog would have a closer relationship with a Basset Hound image than it would to a Bald Eagle image. Therefore within the vector space (the collection of all vectors) the bulldog image would be in close proximity to the basset hound image; this is referred to as Semantic Proximity. This is how when using AI you can prompt AI to “find all available images similar to this one” and get results quickly. Of course, you would want to be more specific in your prompt. Nuxeo and its underlying components can easily store your vector space. Embeddings are very important to AI since machine learning algorithms cannot understand native images, documents, etc. without creating vectors of them. Embeddings therefore, become a key step in AI-enabling Nuxeo; easy, right?

Now that we know that embeddings can easily be stored in Nuxeo, how can we connect Nuxeo to available AI tools/services? One option is to call AI directly to create features within Nuxeo for auto-tagging, dynamic metadata extraction, semantic search and more. And you can even embed a model directly in Nuxeo so that your content/data will remain close and private to your chosen model, no need to send data to the Cloud to leverage a model. In fact, at Maretha, we built a generic addon for basic AI interactions with any model; feel free to take a look at it here maretha-ai on GitHub. Another option is to leverage an AI agent (think of this as a digital, non-physical, robot that can perceive, reason and act on data/content) of your choice wherever you run agents (e.g. - AWS, Google, Vertesia). For your agent to interact with Nuxeo you need a Model Context Protocol Server (MCP server), which is basically a bridge that can connect to any AI agent that supports MCP. There already exists a Nuxeo MCP server exposing Nuxeo REST APIs.

At Maretha we have worked with many clients to help them understand the “art of the possible” to AI-enable Nuxeo. Whether you need to make your content collection more accessible through simple prompts, perform analysis across a subset of your content collection, or streamline a process by first understanding the content ingested and then decide what path it should take, we have heard and seen a lot. My advice to you is simple, do not wait for your organizational leadership to come to you and say “I found a great AI vendor that can AI-enable our DAM/ECM system and I want you to look at it.” Figure out how AI-enabling Nuxeo can help your organization and reach out to us to help you understand how to proceed. By doing so you will be perceived as a visionary within your organization making you a valuable resource; and in the times we are living in, this can be very important.

← All insights