Enhancing Content Management with AI-Driven Tagging

Designing an AI-powered tagging system for content authors, streamlining the process of adding relevant tags and alternative text to enhance efficiency and accessibility.

Problem:

Manual Article Tagging:
Content writers were spending significant time manually tagging articles, which slowed down their workflow and led to inconsistent tagging. This hindered efficient categorization and discoverability, especially as the content library grew.

Alternative Text for Images:
Writers were also strained for time when adding alternative text (alt tags) for images. This step, essential for accessibility and SEO, was often rushed or skipped, impacting the user experience and accessibility standards.

Initial Exploration of Wagtail Platform

Before diving into research, I first explored how Wagtail was currently using tags and alt text within the Federal Reserve's content management system. I reviewed existing workflows to understand the limitations and pain points content authors were facing. This hands-on exploration provided valuable insights into the current state of the platform and helped identify areas for improvement.

By analyzing how tags were manually applied and alt text generated, I pinpointed inefficiencies, especially in terms of time constraints and lack of automation. This exploration laid the groundwork for the research phase, ensuring that the solutions I proposed would directly address these challenges.
In this phase, I also considered how AI could enhance these workflows without disrupting existing practices, ensuring a smooth integration that would improve efficiency and accuracy for content authors.

Research and Discovery

Initiated the project by conducting a competitive analysis to explore existing implementations of generative AI in content creation platforms. The goal was to identify best practices and understand how similar tools had been successfully integrated into workflows. While there were no exact solutions that directly addressed Wagtail’s needs, I uncovered related AI features in other platforms that offered valuable insights. This exploration laid the groundwork for adapting and customizing those concepts to create a solution tailored to Wagtail’s unique requirements.

User as the Pilot:
Ensure user is the pilot. AI should serve as an assistant, giving user full control over decisions and actions.

Validation:
Ensure user has the ability to validate and edit AI-generated tags.

Feedback:
Ensure continuous feedback loop between AI-generated content and user. This will help refine and improve system over time.

Concept development and user flow design

In the conceptualization phase, I mapped out the user flow to cover all potential paths and interactions across two key pages: the Image Upload page and the Article Building page. This ensured that every step in the user journey, from uploading images with alt text and tags to building and tagging articles, was thoroughly accounted for.

For the Image Upload page, I focused on integrating AI to streamline the process of adding alt text and tags. Users would have the ability to accept, reject, or edit AI-generated suggestions, with a smooth option for manual tag additions.

On the Article Building page, the integration of AI-powered tag suggestions was a key feature, while still allowing users the flexibility to manually input or modify tags as needed. This was designed to make the article-building process faster and more efficient while still giving the user full control.

During this phase, I also worked closely with developers to understand the technical constraints and ensure that the design would be feasible to implement. This collaboration helped identify any limitations or potential issues early on, allowing us to make informed decisions about which features could be included and how to best approach the AI integration within the system’s technical framework.

UI design and prototyping

With the design goal because to enhance user experience through integrating AI into existing workflows, I created detailed wireframes and interactive prototypes using Figma.

Here is a flow of one of the pages with design implemented

I didn’t automatically generate tag options in the design; instead, I gave users control over the process. They can choose whether they want to generate tags or text by selecting their preference with a button first, allowing them to take the lead in how they use the feature.

Users can:

•Generate suggested tags using AI.

•Add or remove tags to ensure the final list meets their needs.

•Regenerate suggestions for additional options if needed.

•Provide feedback on the AI-generated tags to improve accuracy and relevance.

Future Considerations

Early feedback and usability testing showed that users found the new feature valuable, as it helped streamline their workflow and save time.

In the future, we plan to automate the process based on AI-generated recommendations. By analyzing user interactions and feedback, we’ll refine the system to provide intelligent, context-aware suggestions while ensuring the experience remains intuitive and user-friendly.