What AI capabilities have been integrated into Drupal?
The recently announced AI Initiative in Drupal is being used to coordinate efforts on a number of fronts. Abstracted layers are being built out so that Drupal is able to integrate with a wide range of systems, including semantic databases and external APIs. The pluggable and configurable nature of Drupal is being extended to handle many new AI capabilities.
Capabilities are being added in many areas:
For end users: Chatbots and semantic search. Unlocking the knowledge contained within SQL databases. No more is knowledge retrieval about keywords. It is now about meaning. This opens the way for a vastly improved user experience.
For content editors: Content can be augmented in a number of ways using LLMs and other APIs. For example, content can be generated or adjusted. Common use cases will be to make corrections for tone of voice or style guide, summarising and idea generation. Editors will need to adapt to new ways of working. They will come to expect content augmentation tools to be built into the CMS.
For ambitious site builders: AI agents can be leveraged to help build out sites. A whole new layer is being built on top of existing Drupal APIs to allow site builders to interact via a prompt. Drupal will collect the required details, and the structure of the site can be manipulated by API integrations. This will disrupt the standard site-building flow for Drupal sites. It is a massive shift in the way sites can be built. Drupal was once organised around editor UI, then on developer configuration, and who knows, maybe agents will be coordinating in the future.
For integrators: It will be possible to orchestrate complex flows. The ECA module is being extended to integrate with the AI ecosystem, making it possible for architects to wire together logic flows which make use of data within and without Drupal. Take a step back and imagine the CMS as a more generic content store - not just for website content. There are a number of possibilities which will open up in this space if Drupal is able to orchestrate across all of these services.
What’s unique about how Drupal handles emerging tech?
Drupal is based on a decentralised structure. The Drupal Association is the coordinating body, and there is a strong band of core contributors and supporting partners on which the project relies on to progress. However, innovation does come from creative individuals who have a passion and their own itch to scratch. This means that innovation is able to occur on many fronts and will tend to favour what is popular or needed at the time.
Most recently, we have seen a number of initiatives where Drupal is innovating in a coordinated way. Recently, we have seen strong gains made in Drupal CMS, Recipes, Experience Builder, the Marketplace, Project Browser and Automatic Updates and of course the AI initiative. The world is changing quickly, and you are either evolving or dying.
Are clients asking for AI? If so, what for?
With our clients, we have seen an interesting dynamic. The initial responses to AI was one of amazement and excitement… and then a little fear. People tend to divide into a number of camps with how they view AI: the sceptics and the promoters. There is uncertainty on how it will all play out. There are possibilities, but they need to be proven. As things have progressed, it is becoming more acceptable to incorporate AI into workflows for increased efficiency.
In the Australian government, there has been a natural concern for how AI is being deployed and utilised. It is a new area which has been unregulated. Governments, like all organisations, are concerned with the accuracy of responses and the privacy of data. This goes to trust, security and integrity. Once the safeguards are in place, AI implementations become easier to implement. We are now seeing governments willing to adopt AI into their projects and to utilise the services of suppliers.
We have seen clients most interested in the following areas:
Content personalisation via AI-powered recommendations: Machine learning with user behaviour and content analysis can provide superior recommendations to users.
Content augmentation: Owners of large corpuses can leverage the power of AI for metadata extraction and tagging of images. Tasks which would have been impossible to do manually are now achievable.
Chatbots: Organisations can unlock their content by adding chatbots, which are able to utilise content embeddings to return more realistic results.
Semantic search: Users expect to interact with search in a more conversational way. It is now no longer enough to be based on keywords alone. A new generation of search will evolve with this new way of interacting.
How do you stay ahead of trends in open source?
How do you stay ahead? That is a life question.
- Be curious.
- Have the energy and the courage to forge your own path.
- Join the dots and then extrapolate.
- Apply yourself to the problem and stand on the shoulders of giants.
These characteristics sum up healthy open source communities.
Can you share an innovative use case?
Tone of voice
One practical example is how LLMS can be used to make recommendations to improve the style or tone of a piece of content. An LLM can be fed a number of rules that should be adopted for a certain organisation. For example, the use of abbreviations and slang, the tone of voices, the level of difficulty, and the length of sentences. These rules can then be applied.
I like this as an example because AI is being used to support the user in delivering the content in a way that supports the organisation's goals. It is not about replacing the author, but rather supporting them.
Semantic search
The area I am most excited about would be the addition of semantic search capabilities to Drupal websites. This opens up a lot of possibilities for content owners, such that the true value of content can be unlocked. Semantic search allows for similar, but otherwise removed, items to be retrieved and combined to produce meaningful responses and potentially new insights. This extends the core functionality of Drupal as a CMS and is a big win as it will become more widely adopted.
For example, we have adopted semantic search on GovFlix to return better results for users as well as to provide a chatbot interface into the content..
What makes Drupal future-proof compared to proprietary platforms?
I don’t think that it is possible to be 100% future-proof.
Take a look at the composition of a stock market index over the years. Successful companies, those who were leaders, eventually fall to the side and are over taken. This rise and fall can largely be attributed to a changing environment and an inability to adapt. Technology is the main catalyst here. New developments can instantly wipe away the edge of the incumbents,
Looking back over the past 25 years, there has been a huge amount of change in the web space. Back in the day, the ability to “Edit this page” was a groundbreaking concept. Drupal has ridden the successive waves of change since these early days. Frameworks such as Drupal and others need to be able to adapt to the new technologies that come along.
I would look at it as a balancing act. In order to be future-proof, you need to be stable and adaptive. Borrowing the concept of "antifragility" from Nassim Nicholas Taleb:
"Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the antifragile gets better."
https://en.wikipedia.org/wiki/Antifragility
Drupal has been forced to be antifragile to survive. Over the years, it has faced a number of technological, cultural and organisational challenges. It has had to solve a number of problems and adapt systems to manage this change. As a project, it has survived and innovated through each of these challenges.
In summary, choosing open technologies with active communities is one way to maximise the odds of being future-proof. Choosing technologies that are able to innovate with new ways of technology is an extra step in this direction.
How quickly can you integrate new tools or APIs?
Successful projects are the result of many factors that require alignment across many fronts. Technology is not the only determinant of speed. There needs to be the vision, the commitment and the capability in place to deliver a successful project.
On the technical delivery side of things, there are a few factors at play:
- Teams with know-how and practice experience
- Access to tools and components when suitable
- Ability to customise if needed
- Well-tested quality and deployment processes.
The Drupal ecosystem shines when it comes to addressing each of these concerns. There is considerable breadth and depth of modules and suppliers available to assist with the integration of tools and APIs. Drupal has a solid plug-in system, which does make it easily extensible for new requirements that may come along.
Read the full article on DrupalSouth: Open Source and Innovation: How Drupal is Keeping Pace
