Companies often have a large amount of free text content and documents that contain valuable insights but remain unused due to the sheer volume and complexity. For example, our client in the pharmaceutical healthcare sector was collecting customer data in their CRM system, including free text feedback from customers and sales reps. However, analysing this data and incorporating it into strategic assessments and decisions was often a challenge.
The system we developed solves this challenge by using Large Language Models (LLMs), which classify free text content into predefined content classes. Thanks to the automation of this refinement process, unstructured information can now be efficiently converted into structured data. The derived tags serve as inputs for customer journey tools, providing valuable insights and facilitating decision-making for sales strategies. In short, our system makes free text data analysable.
Advantages of the tool:
Our solution is based on Azure OpenAI, GPT models, LangChain and Python and is operationalised in Palantir Foundry. It is more performant than previous solutions based on NLP transformer models.
Our solution for automated tagging makes it possible to efficiently categorise unused free text content and thus improve the customer's database. The scalable approach ensures high processing speed and accuracy when analysing large volumes of data.