Blogs

Beyond content chaos, clarity with entities

By Vincent De Geyndt posted 07-19-2024 02:49

  

Content Chaos, the natural state of information

Putting content under collective control is a growing challenge. With the adoption of cloud content services, the number of content systems in organizations continues to increase, including shared drives, sites, and email systems. With the migration of content to Sharepoint, small-scale collaboration is empowered. But governance, compliance, and collective clarity are still lagging behind. Where can I find my document? What were our agreements back then? Content chaos thrives.

The reality of agility is that it is difficult to adapt content systems to a changing business environment. Running document archives and team collaboration tools eats into tight IT budgets. Regrettably, the primary innovations that we introduce into ECM and content services are merely enhancements. Nobody, except for startups, has sufficient time to leverage generative AI and exciting new technologies.

On the other hand, your employees and customers expect to access their content 24/7 on demand, enjoy fully digitized content centric processes, and rigorous protection of their confidential and private information.

Therefore, we need a new way to organize our content that goes beyond websites, folders, types, and groups. Duplicating, tripling, or multiplying your customer's info in content metadata is a red flag for your content platform.

Unlock Clarity!

In the past decades, there haven't been any really innovative approaches to context modeling practices. New and established players have recycled old concepts like folder structures, document types, and categories. They invented new weapons of mass distraction, such as sites and drives. Although useful for short-term, small team collaboration scenarios, they are not helpful for long-term collective information management. When I helped roll out Documentum at BASF in 1999, it had the same organizational principles as most ECM's today. Around 2005, open-source ECM software, like Alfresco and Nuxeo, became more flexible. Introducing aspects and the use of NoSQL databases to scale up expanding content repositories in document models. But we didn't see any significant changes in ECM.

Since then, we've noticed a difference between managing content with SharePoint or Microsoft 365 and the rest (the archive or repository). Online collaboration is enhanced by the ability to edit a document in parallel. The plates of collaboration, transactional content services, and records management have gotten separated.

Today, we think a new way of sharing content is needed to solve the chaos. The problem is, first and foremost, too much information, and more folders won't help. The way we model the context should be reconsidered to keep content. Using ideas from other fields can help us a lot.

Most 'creative' content produced by (collaborating) humans becomes static after a short, intense, and creative period. Consider a sales proposal, an R&D design, product documentation, commercial presentations, and other similar documents. After approval and exchange, the document is cast in stone and considered a record.

Document types are fuzzy

Content is very heterogeneous by nature. A document is a snapshot of information in time. If you agree with this definition, it's no surprise that typifying documents is hard. Unfortunately, classes (types) do not align very well with the heterogeneous nature of many documents. They do match well with transactional documents, such as a proposal or purchase order. Because anyone can add anything they want. Of course, we have a set of transactional documents (a quote, a contract, an invoice, a deliverable, …) that are closely related to well-defined processes. They align well with types. However, there is also the rest. Documents that are of type “document”.

A collection of groundbreaking research titled “Researchers at the University of Cambridge propose Anchor AL: a unique machine learning method for active learning in unbalanced classification tasks” shows this problem of unbalanced classification. It's human to want AI to clean up the mess, but it's better to avoid it. Common sense is not outdated yet.

All of these rest documents are tied to rights that matter to your business. The ContentGrid platform allows for a lot of different classification combinations without the need for manual classification. However, it provides the opportunity to categorize a document into a specific type, such as a quote, with an amount, a product, and a customer, by introducing associations.

We had a problem in the past where we couldn't use multiple inheritance because we couldn't make a document both a project document and a meeting note. Thankfully, Alfresco introduced aspects to model this challenge. ContentGrid takes it a step further by reinstating entities and relationships as the optimal modeling mechanism, given its extensive and mature toolset.

Have we already mentioned our fondness for postures? The ecosystem is 49 years old and stronger than ever. Expect Postgres-based technology to make waves in analytical user cases in the future. ContentGrid will happily integrate to make the search experience even better.

Unlock clarity in Content Chaos

ContentGrid alone can't bring order to content chaos, but we think it's the best way to help our clients, partners, and co-creators understand their content processes and repositories, and rethink how they create context for content. Finally, ContentGrid will allow customers to learn together without compromising privacy or confidentiality. This exciting new development, captured in the research domain of federated learning, will allow organizations to work together in domains that they gain from collaboration. All insurance companies benefit from detecting fraud, so why not use AI to learn together to make the world a better place? If HR companies or divisions of an HR group collaborate to enhance the matching of candidates and job offerings, we can enhance the business. If ContentGrid can contribute to a greener world, but a simple and resource-conscious architecture, and maybe help to federated learning, we can support the only way to tackle the big challenges our planet faces in the future: by working together.

Read all about it in out Ebook

0 comments
7 views

Permalink