TerminusDB Logo all white svg
VectorLink - The TerminusDB Semantic Indexer - Logo White

Semantic AI Classification

LLMs provide powerful tools to leverage free-text and unstructured data. We can help you programmatically solve categorisation and taxonomy headaches.

How Can We Help

We have built a semantic indexer called VectorLink which, when combined with our data engineering knowledge, enables us to solve traditionally difficult problems using LLMs.

Classification tasks on unstructured data are difficult, but LLMs give us powerful tools to help us leverage free-text and unstructured data. Our services enable you to supplement generative AI’s parameterised knowledge with your content in a way that is meaningful to the LLM. Tasks that currently take thousands of hours to manually complete can be programmatically solved in a vastly more efficient and cost-effective way.

Semantic AI Classification Use Cases

Taxonomy Creation

Unstructured corpora can be given taxonomies to help content be more discoverable and easier to navigate. Research, medicine, libraries, legal and other sectors that rely heavily on large amounts of unstructured data can improve efficiency and achieve their goals quicker by spending less time digging for key information.

Automatic Semantic Tagging and Categorisation

We can assign categories or tags automatically to complex content, freeing humans from the loop. One such example on the OpenAI Forum highlights this need. The user, unable to programmatically categorise unstructured data, has manually had to classify 2 million products for a supermarket chain.

Audience Creation

With a set of personal profiles derived from an individual’s content and a small training set of known outcomes (e.g. purchasing behaviours), we can construct audiences tailored to specific goals. Whether it is product sales, cross-selling, brand advocacy or more complex goals, using an LLM’s parameterised knowledge with what you know about your audiences will improve your desired outcomes.