FIF™ uses the power of AI, natural language processing APIs, and deep learning algorithms to create custom classifiers and classes to index and catalog content rapidly. Using APIs such as IBM Watson’s Visual Recognition, FIF™ helps distinguish not only what a fish is, but also what species and subspecies it is, such as a Bigmouth Buffalo, as an example. Thereby we follow one of the first significant benefits of FIF™, Speed of Experience. By making the application familiar and obvious from the start, content editors do not waste time with learning a new workflow, they instead use an intuitive application. Because machine learning is a core feature of FIF™, the tool can be trained to more accurately identify and tag content over time as it’s fed more and more assets. Thereby minimizing end-user burden from classifying on their own to solely validating the machine-learning-driven results.
FIF™ creates that one thing that is missing from sites today, a "Just for Me" experience.
Find-It-First™ Case Study: Department of Energy, Office of Science (SC)
agencyQ has been a long-time partner of the Department of Energy, Office of Science (SC), dating back to a Sitecore site redesign and mobile optimized site build in 2011. We expanded our support relationship in 2015 when DOE SC approached us with the challenge of future-proofing over 11,000 pieces of content that serve varying stakeholders, including scientists, politicians and members of the media.
Understanding DOE SC’s challenges, agencyQ proposed an innovative approach of building a content repository using Coveo—an industry-leading search engine—and IBM’s Watson Natural Language Understanding API to gain insight from structured and unstructured content. We developed Find-It-First™ (FIF™), an unprecedented, strategic and technical approach to managing content and assets on a content-rich website. Built to create both a quicker and more fulfilling user experience, FIF™ emphasizes speed, accuracy, and flexibility by leveraging existing technologies and historical analytics to enrich data for warehousing into a central content repository. Using the API enables us to influence search results based on the user’s likely persona defined in Sitecore. Resulting in content that is more accessible, reliable and relevant to users. Thereby enabling them to interact with the site in a manner more natural to them, all while reducing the burden on the content editor. Additionally, initial tests have shown significant performance improvements with the page load speed scores going from the low 40s to low 90s, based on Google’s Page Speed Insights.
We continue to enhance the content repository by ingesting additional data sources from within SC, as well as creating analytics dashboards that show content editors how particular pieces of content are performing and user search and engagement patterns.