Kickstarting semantic AI projects in Tridion
22 Nov 2022
5 mins
We recently hosted a webinar with leaders from our partners the Semantic Web Company (SWC) and WAND, who, alongside the Tridion team discussed how Pool Party’s (a SWC product) taxonomies can be kick start semantic AI projects in Tridion Sites and Tridion Docs.
Florian Bauer, CSO (Chief Solutions Officer) at SWC and Ross Leher CEO at WAND, led the presentation. The conversation kicked off with an overview of the issues facing companies. According to Florian, finding content is one of the most significant issues that companies face today, leading to poor user experiences and negative sentiment. Did you know that “65% of total content creation investments are wasted due to content not being found or used by the organization?”
In simple terms, taxonomy is a list of terms used to categorize and find information without browsing through every file, image, document, or web page. Using a taxonomy allows users to find and reuse content. Well, it is no surprise that taxonomy is one of the most vital components of building a comprehensive and effective content strategy!
Simple as it may sound, building taxonomies is no easy task; in fact, it is one of the most complex and time-intensive tasks. It involves creating, applying, and maintaining a taxonomy of business information. It needs to consider all the content that has been accumulated and will accumulate in the future. With ever-increasing content volume, this is becoming even more complex.
Taxonomy’s structure should capture the relationships between the information in an intuitive way. It needs to be updated regularly to remain relevant and useful as new content is added and changes occur in terminology, technology, and markets.
To complicate things further, information is multi-dimensional; users will access it from different departments or teams for various purposes. It isn’t easy to account for all possible ways of tagging so that these different users can easily find content.
Building an effective taxonomy requires an intimate understanding of the users and use cases and mapping them with the content, which comes over time and experience. An effective taxonomy thus can’t be built in a short period without deep knowledge and expertise.
Need for an effective taxonomy tool in Tridion
Florian mentioned that many of SWC’s customers came back to understand the amount of time it takes to build taxonomies and ultimately start using semantic AI. Taxonomies is the first and time-intensive process needed for smart tagging and hence for semantic search to work. Intending to expedite the taxonomies process, SWC partnered with WAND, which provides an exhaustive list of expertly curated taxonomies that span 35 industry verticals and every function of the enterprise (core, IT, finance, marketing, HR, project management, fleet management, etc).
Build effective taxonomies and save significant time
Ross Leher from WAND explained how these expert-curated taxonomies can be imported into Pool Party and Tridion. He mentioned that WAND’s taxonomies are vast and exhaustive because these have been built since 1983. Starting with insurance companies, the scope was expanded to include the 35 verticals it covers today.
WAND’s taxonomies repositories help clients accelerate sales by 6-12 months, expedite client purchase decisions, create effective demonstrations for clients, and enable quick deployment. WAND’s taxonomies are selected and delivered online and are formatted for direct import into enterprise applications. The whole set-up is completed in a matter of a few minutes.
Importing taxonomies from WAND requires only two steps to import the taxonomies into the workflow and incorporate them into the Tridion architecture. It has the foundation metadata module to inform the AI of the curated structure that the AI engine can refer to. Usually, it takes 6-10 weeks to create a metadata model or a given domain of knowledge. With WAND’s curated taxonomies, this can be reduced to 7 to 10 days. Thus, speeding up the semantic AI implementation significantly.