All Eyes on AI at DrupalCon Lille

Whether you’re a skeptic or an enthusiast, artificial intelligence (AI) is already integrated into your everyday lives. From voice-to-text converters, digital assistants, and product recommendations to customer service chatbots, you’ll find AI under the hood of many tools and services you interact with on a routine basis. 

Here at DrupalCon Lille, AI has also been ingrained in daily conversations. There have been 15-minute coffee break presentations, including Dropsolid’s live-demo of AI within Drupal CMS and overview of the possibilities of AI in open source Mautic, as well as FFW’s review of the state of AI in Drupal. (Promising news: There are 168 modules available that implement some sort of AI-related functionality, of which 77 are compatible with Drupal 10 already!) And, at the Acquia booth, we offered attendees the opportunity to chat with ChatDriesPT, an AI video chatbot that answers your questions about Drupal 10.

Image

Of course, there have been numerous sessions dedicated to AI, but even in sessions that aren’t explicitly about the technology, it’s regularly raised as the Drupal community grapples with the question of how AI will impact the future of Drupal and the digital future overall. Read on for a taste of the AI conversations happening at DrupalCon.

Unlocking the potential of digital experiences with Drupal and AI

We heard about the magic of AI and how we can use it to deliver captivating end user experiences from Saranya Rajaram, Business Analyst from Axelerant, an Acquia community partner. Given the ability of AI to process vast amounts of data, learn patterns, and make intelligent decisions, it can revolutionize the way businesses engage with customers, especially when combined with a digital experience platform (DXP) based on Drupal.

As businesses seek innovative ways to elevate their online presence to create experiences that meet customer expectations and generate higher satisfaction and loyalty, it only makes sense that AI and DXPs come together. AI will transform Drupal into an intelligent, adaptive platform capable of delivering highly personalized, seamless, and engaging experiences. Saranya highlighted five enhancements to DXPs that AI makes possible:

• Personalization and recommendations: To enhance user engagement and satisfaction by delivering content and experiences that align with the user's interests and needs
• Conversation interfaces: To make human-computer interactions feel more like human-to-human conversations, which enhances the user experience
• Automated content management: To enhance the content management process and overall user experience by improving content quality and providing a more personalized and efficient content experience
• Enhanced search: To create a more user-centric, intelligent, and engaging search experience that leads to higher user satisfaction, increased user retention, and improved content discoverability
• Data analytics and insights: To optimize the digital experience, engage users effectively, and achieve business objectives through data-driven insights and actions
   

Besides recommending various tools for each of these possibilities, Saranya also shared some considerations when it comes to best practices for integrating AI tools with Drupal:

• Detailed requirements analysis
• Choosing the right AI tools
• Your technical expertise
• Data preparation and quality
• Data privacy and compliance
• Maintenance and updates
• Performance optimization
• Interpretability and explainability
• Diversity and inclusion
• User training and adoption

Saranya also reminded session attendees to recall the limits and challenges of AI tools, many of which overlapped with topicsDigitalist covered in their session.

Integrating AI as a new colleague

In a packed session, Ulrika Mengshoel Wedelin, CEO of Digitalist, and David Holmlund, Chief AI Officer at Digitalist, shared their journey of integrating AI tools as a member of their team. Beginning in March, their process for onboarding their new AI coworker started with a management team decision to pay all employees’ licenses for AI tools. The agency created an AI policy around ethics and data privacy, and established the chief AI officer role. They hosted meetups around AI and included a section dedicated to discussing AI during every staff meeting. Digitalist’s motto became, “Can AI help me with this?”

The new additions to the Digitalist team included: 

• A 24/7 assistant to help with content creation, idea brainstorming, feedback, and domain knowledge
• A tireless developer to help with code by suggesting, editing, streamlining, or explaining lines of code
• A media wizard to help with creating image, video, music, and speech content
   

With each of these new “colleagues,” Digitalist netted both gains and pains. While they offered quick ways to get started and solutions for procrastination, trade-offs included the need for human oversight, information security, ethical concerns, and the work-in-progress nature of these tools —  AI still hasn’t figured out how to illustrate anatomically correct hands, for example. Ulrika and David outlined the concerns they weighed as they adopted these AI tools for Digitalist: ethics, bias, impact on jobs, competence, laws and regulations, fraud, transparency, AI hallucinations, corporate use, emergency consumption, personal data and information security, and much more.

Ultimately, they advised other companies looking to empower employees to start using AI tools to:

• Establish guidelines
• Have continuous awareness trainings
• Be open to learning
• Be compassionate with insecurities amongst colleagues
• Discuss, discuss, discuss
• Just start
• Evaluate, learn, and discuss some more!

Redefining content creation workflows with AI

The excitement and anxiety around AI didn’t stop Chris Wu, Senior Web Developer at Tern Bicycles, from ideating and creating new ways to leverage AI as part of the company’s web content workflows. While there are many types of AI models, Chris focused on large language models (LLMs) and their four key capabilities:

• Expand: Generate useful and high-quality content quickly
• Summarize: Extract a summary of long-form content while preserving the most important information
• Transform: Translate, convert, or alter tone of your content to match your brand or audience
• Inference: Analyze your content and identify hidden patterns or insights to optimize your workflow
   

Chris highlighted the potential of using AI throughout the content creation process, including existing capabilities from Drupal modules, some of his organization’s own contributions, as well as ideas for future development:

• Content drafting: AI can be used to produce an initial draft of content, with the Open AI: CKEditor Integration module offering features to complete text, adjust its tone or voice, summarize, translate, or reformat.
• Content review: To better understand personas and refine content, Chris suggested using AI tools to help with readability and proofreading. You can define your AI model with a persona and have it score your content’s compatibility with the needs of that audience, or you can check writing style against guides like AP Style or AMA, or even check language variations, like U.K. vs. U.S. spellings.
• Translate: Without losing context, tone, or voice, AI translations are beginning to outperform neural machine translations like MS-Translator. Chris suggested the combination of the Drupal module TMGMT with OpenAI Translator, his company’s contribution, to manage all translations and use OpenAI as the translator.
• SEO: AI-driven SEO tools can provide deep insights into what audiences are looking for and help with metadata creation, such as suggesting SEO-friendly titles or keywords. 
• Campaigns and social media: AI can imitate tone and understand each social media platform’s restrictions, so it can be used to create posts. Chris shared an idea for an AI Readability Toolbox that would leverage predefined personas to create targeted social media posts.
   

As we see from this session and the others, there are many different ways to leverage AI. While AI has numerous well-known challenges and limitations, it also has countless benefits. All eyes are on AI. It remains to be seen what the Drupal community does next when it comes to defining the future of AI with Drupal, but for now, Dries Buytaert has ideas about what may be around the corner — learn what the creator of Drupal believes we can look forward to with AI and the open source platform.