FUSE Topic Modeling™: AI-Powered Content Classification

You may have a list of taxonomy (e.g. topics, classifications, tags) that are meaningful to your users. Many organizations, however, are stuck: years and years worth of valuable content with varying degrees of classification applied, inconsistent or different classification across content sources, or no classifications at all.

If you have hundreds, thousands, hundreds of thousands, or more of unclassified documents residing on one or more of your content sources, then FUSE Topic Modeling™, powered by state-of-the-art machine learning models, is for you.

How It Works

In short, step one is to train the topic modeling engine with some sample content that is already classified to your specification. Once this is done, FUSE Topic Modeling™ (included with all Enterprise, available as an add-on to Standard, engagements) takes over. After your content gets indexed via one of our many integrations, the topic modeling engine runs and assigns classifications per your taxonomy model.

Frequency

You can run the topic modeling process one-time — helpful in situations where you need to classify historical content and you plan to self-classify moving forward — or you can run it continuously taking the pressure off of content editors to classify content correctly.

Usage Fees

Typically content only needs to be modeled once. So you may have 1GB of content to be modeled but that will only be a one-time modeling event. Thereafter, you may only add, for example, 5MB or so a week. As a point of reference, about 500 pages of text (2,000 characters of text per page) equals 1MB which would trigger a $2.50 usage fee (yes, two dollars and fifty cents). Usage fees are billed quarterly.