Fluid Topics v5 background

Release notes

Fluid Topics Version 5

We are excited to announce the release of Fluid Topics – the v5. This new version will bring many improvements and further empower existing users with core functionalities such as enhanced search capabilities, generative AI and increased performance.

Major Upgrade to Our Tech Stack

Over the past two years, the Fluid Topics team has worked tirelessly on overhauling the content processing back-end. Today, we’re delighted to introduce a new system designed to enhance the scalability, speed, and security of our platform. In the coming weeks, this significant upgrade will also introduce:

  • a completely new module to define Access Rules
  • new features in the Publishing screen (and more!)

After months of development and refinement based on client feedback, the Designers have now reached maturity and are production-grade, transitioning from Beta to Stable. Having been successfully used in several major projects, they have become the preferred method for creating the user experience and branding you desire.

The Analytics stack has also received significant upgrades: events are now processed in real-time, eliminating asynchronous daily processing. It allows for immediate retrieval of portal events as they occur and provides real-time updates to dashboards. With the installation of new hardware, our data lake now responds faster, significantly reducing the response time of our web services and screens. Leveraging this enhanced capability, we now offer more data than ever, with up to nineteen months of data retention across all our analytics.

Now let’s talk more in depth about Generative AI, Page Designer, Admin and Content Processing improvements!

Generative AI

It’s no secret! We’ve been working hard on integrating Generative AI capabilities into the Fluid Topics platform. As AI is a constantly evolving field, we have an ambitious roadmap to provide you with readily available, tunable, and controlled AI that delivers true business value. We can now officially introduce our new Generative AI module.

With version 5, our GenAI add-on is already well-established, with more features coming soon. Here are some of the key highlights:

Semantic search:

  • Improved architecture to provide better service availability.
  • Complete semantic indexing (aka embeddings computation) of both structured and unstructured content, including PDF documents.
  • A new version of the semantic search API enabling searches in all languages is now available.

Large Language Model (LLM) completion:

  • Automatic test of the LLM connection within the GenAI profile configuration screen.
  • Support for additional externally hosted LLM models and introduced a self-hosted LLM now available for alpha testing.
  • Dedicated analytics screen and related web service for comprehensive visualization of GenAI query usage per profile.

RAG-powered Chatbot

In version 5, Fluid Topics introduces a new RAG (Retrieval-Augmented Generation) chatbot to enhance user interaction and support. This advanced conversational bot uses semantic search capabilities alongside generative AI to deliver highly accurate and contextually relevant responses.

By incorporating this RAG chatbot, we aim to provide users with a more intuitive and efficient method for accessing and engaging with their content. It will be like talking to a human!

This addition underscores our commitment to advancing user support and enriching the overall platform experience.

RAG Chatbot Fluid Topics

Additional GenAI-related capabilities:

  • New use case examples for the search page, including a summary of search results and keyword suggestions
  • New case deflection component based on GenAI features
  • Managing the selection of filters on question-answering pages

Page Designer

Our mission has always been to empower our clients to derive the maximum value from their product knowledge and from their documentation portal. In addition to being promoted to stable, the designer introduces new capabilities for all designers.

The Reader Designer has been significantly enriched with the following features:

  • New components: Switch to a document’s relatives, Search results, Search in doc with highlight, Attachments, and more.
  • New Action blocks including the ability to send feedback

We have also introduced several other enhancements, including improvements to the table of contents and code blocks with a copy-to-clipboard feature.

As for the Search Results Designer, it now offers more possibilities:

  • The ability to cluster results,
  • The highlighting of missing keywords, to help users better understand their search results,
  • The addition of new Action blocks, such as “download unstructured document”.

But that’s not all. For an even simpler setup, v5 includes a gallery of examples and default templates. You’ll find ready-made layouts for Homepage, Search results, and Reader pages, all with existing components pre-arranged. You can easily customize these pages to fit your visual identity.

Admin Improvements

As always, we strive to increase administrators’ autonomy in managing advanced portal configurations.

To support this, v5 introduces new administrative features:

Configuration history: This feature allows admins to inspect the latest configuration changes on a portal.

Configuration history feature

Confidentiality with watermarking: this feature lets you configure dynamic watermarking of documents based on rules, or to prevent users from adding confidential documents to their personal books.

Content Processing

We are thrilled to share that our latest version now supports DITA-OT 4.x, offering enhanced functionality and better performance. In addition to this integration, we’ve improved our Paligo connector, particularly for glossary management, and made updates to our Markdown, Word, and Madcap connectors.

We’ve also added Word, Markdown, and OpenAPI as default sources, simplifying your content management workflow.

Version 5 comes with new features such as ft:publishStatus and publication visibility scheduling capabilities (ft:searchableFrom).

Furthermore, our Enrich & Clean feature has been upgraded with new semantic metadata, enabling even more accurate content enrichment and cleaning.

Please note that the Control File Legacy will be deprecated from the UD connector starting from version 5.1; please contact our support department if you have any questions.

For more information, check the full release note

Please give this new version a spin
and let us know what you think.

If you want to know the roadmap and participate to the choices of the next features, feel free to contact us.

We develop Fluid Topics for you.