Problem

The information architecture of HHS.gov faced significant challenges in ensuring consistency and alignment with the context and intent behind both external and internal user search queries. The existing manual processes were lacking quantitative inputs, time intensive and inefficient, leading to analyst-injected bias, and inconsistent categorization across a large volume of content. The process lacked a strategic project plan and execution strategy, resulting in fragmented content management.

Solution

Analytics Logic was engaged to develop a Natural Language Processing (NLP) solution tailored to enhance the information architecture of HHS.gov. The project involved several key phases:

1. NLP Project Plan Development: Analytics Logic crafted a detailed project plan for the NLP implementation, aiming to drive the development of entity topic-based information architecture. This plan included milestones for technology integration, training, and deployment phases.
2. MVP Creation and Deployment: A Minimum Viable Product (MVP) for the NLP solution was developed to demonstrate its efficacy in categorizing content, scoring it for relationships and interlinking and aligning it with Google’s NLP API entities. This MVP allowed HHS to see early benefits in content alignment and provided a foundation for further refinement.
3. NLP Training and Application: The NLP model was trained using a large dataset of HHS.gov’s content, enabling it to learn and predict categorization based on the context and semantics of the content. Analytics Logic also provided comprehensive training to the HHS content team on utilizing NLP data to improve content accessibility, interlinking and SEO.
4. Integration and Automation: The NLP solution was fully integrated into HHS.gov’s Drupal content management system, automating the categorization process and ensuring that new content was automatically classified according to the new topic-based architecture.

Impact

The implementation of NLP significantly transformed the information architecture at HHS.gov:

• Improved Content Consistency: The NLP system provided a consistent method for classifying and organizing content, significantly reducing manual efforts and bias.
• Enhanced SEO Performance: By aligning content categorization with Google’s search algorithms, the NLP solution improved the visibility of HHS.gov content in search engine results.
• Increased Operational Efficiency: Automating the content categorization process allowed the HHS team to focus on content creation and strategy rather than manual classification, saving time and resources.

Challenges Overcome

Implementing NLP involved overcoming several challenges, including the adaptation of existing digital infrastructure to support AI and machine learning technologies, training staff to leverage new tools effectively, and ensuring data privacy and security in handling sensitive health information.

Tools Used

• Google Natural Language Processing API
• Drupal CMS
• Data analytics and visualization tools such as Tableau for monitoring NLP performance

Conclusion

The NLP project at HHS.gov is a testament to how artificial intelligence can revolutionize information architecture in large organizations. By leveraging NLP, HHS.gov was able to enhance its digital presence, improve user experience, and ensure content met both user needs and compliance requirements effectively.

Analytics Logic’s expertise in NLP and project execution ensured that the project not only met but exceeded the expectations of HHS, setting a benchmark for future AI implementations in government digital strategies.